Promulgation of Air Quality Implementation Plans; Arizona; Regional Haze and Interstate Visibility Transport Federal Implementation Plan

Federal RegisterFeb 18, 2014

Ask Donna

What actually matters in this document.

Text

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 51

[EPA-R09-OAR-2013-0588, FRL-9906-30-Region 9]

Promulgation of Air Quality Implementation Plans; Arizona; Regional Haze and Interstate Visibility Transport Federal Implementation Plan

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Proposed rule.

SUMMARY:

This proposed Federal Implementation Plan (FIP) addresses the requirements of the Regional Haze Rule (RHR) and interstate visibility transport for the disapproved portions of Arizona's Regional Haze (RH) State Implementation Plan (SIP) as described in our final rule published on July 30, 2013. Our final rule on Arizona's RH SIP partially approved and partially disapproved the State's plan to implement the regional haze program for the first planning period. Today's proposed rule addresses the RHR's requirements for Best Available Retrofit Technology (BART), Reasonable Progress Goals (RPGs) and Long-term Strategy (LTS) as well as the interstate visibility transport requirements for pollutants that affect visibility in Arizona's 12 Class I areas as well as areas in nearby states. The BART sources addressed in this proposed FIP are Tucson Electric Power (TEP) Sundt Generating Station Unit 4, Lhoist Nelson Lime Plant Kilns 1 and 2, ASARCO Incorporated Hayden Smelter, and Freeport-McMoran Inc. (FMMI) Miami Smelter. The sources with proposed controls for reasonable progress are the Phoenix Cement Clarkdale Plant and the CalPortland Cement Rillito Plant.

DATES:

Written comments must be submitted to the designated contact at the address in the General Information section of

SUPPLEMENTARY INFORMATION

on or before March 31, 2014.

ADDRESSES:

See the General Information section of

SUPPLEMENTARY INFORMATION

for further instructions on where and how to learn more about this proposal, attend a public hearing, or submit comments.

FOR FURTHER INFORMATION CONTACT:

Thomas Webb, U.S. EPA, Region 9, Planning Office, Air Division, Air, 75 Hawthorne Street, San Francisco, CA 94105. Thomas Webb may be reached at telephone number (415) 947-4139 and via electronic mail at

r9azreghaze@epa.gov

SUPPLEMENTARY INFORMATION:

Table of Contents

I. General Information

A. Definitions

B. Docket

C. Instructions for Submitting Comments to EPA

D. Submitting Confidential Business Information

E. Tips for Preparing Your Comments

F. Public Hearings

II. Proposed Actions Background and Overview

A. Background

B. Regional Haze

C. Interstate Transport of Pollutants That Affect Visibility

III. Review of State and EPA Actions on Regional Haze

A. EPA's Schedule To Act on Arizona's RH SIP

B. History of State Submittals and EPA Actions

C. EPA's Authority To Promulgate a FIP

IV. EPA's BART Process

A. BART Factors

B. Visibility Analysis

C. Explanation of Visibility Tables

V. EPA's Proposed BART Analyses and Determinations

A. Sundt Generating Station Unit 4

B. Nelson Lime Plant Kilns 1 and 2

C. Hayden Smelter

D. Miami Smelter

VI. EPA's Proposed Reasonable Progress Analyses and Determinations

A. Reasonable Progress Analysis of Point Sources for NO

B. Reasonable Progress Analysis of Area Sources for NO

and SO

C. Reasonable Progress Goals

D. Meeting the Uniform Rate of Progress

VII. EPA's Proposed Long-Term Strategy Supplement

A. Emission Reductions for Out-of-State Class I Areas

B. Emissions Limitations and Schedules for Compliance To Achieve RPGs

C. Enforceability of Emissions Limitations and Control Measures

D. Proposed Partial LTS FIP

VIII. EPA's Proposal for Interstate Transport

IX. Summary of Proposed Actions

A. Regional Haze

B. Interstate Transport

X. Statutory and Executive Order Reviews

A. Executive Order 12866: Regulatory Planning and Review

B. Paperwork Reduction Act

C. Regulatory Flexibility Act

D. Unfunded Mandates Reform Act

E. Executive Order 13132: Federalism

F. Executive Order 13175: Consultation and Coordination With Indian Tribal Governments

G. Executive Order 13045: Protection of Children From Environmental Health Risks and Safety Risks

H. Executive Order 13211: Actions That Significantly Affect Energy Supply, Distribution, or Use

I. National Technology Transfer and Advancement Act

J. Executive Order 12898: Federal Actions To Address Environmental Justice in Minority Populations and Low-Income Populations

K. Congressional Review Act

I. General Information

A. Definitions

(1) The words or initials

Act

CAA

mean or refer to the Clean Air Act, unless the context indicates otherwise.

(2) The initials

ADEQ

mean or refer to the Arizona Department of Environmental Quality.

(3) The words

Arizona

and

State

mean the State of Arizona.

(4) The initials

BACT

mean or refer to Best Available Control Technology.

(5) The initials

BART

mean or refer to Best Available Retrofit Technology.

(6) The initials

BOD

mean or refer to boiler operating day.

(7) The term

Class I area

refers to a mandatory Class I Federal area.

(8) The initials

CEMS

refers to continuous emission monitoring system or systems.

(9) The initials

mean or refer to deciview, a measure of visual range.

(10) The words

EPA, we,

our

mean or refer to the United States Environmental Protection Agency.

(11) The initials

FGD

mean or refer to flue gas desulfurization.

(12) The initials

FIP

mean or refer to Federal Implementation Plan.

(13) The initials

FLM

mean or refer to Federal Land Managers.

(14) The initials

IMPROVE

mean or refer to Interagency Monitoring of Protected Visual Environments monitoring network.

(15) The initials

IPM

mean or refer to Integrated Planning Model.

(16) The initials lb/MMBtu mean or refer to pounds per one million British thermal units.

(17) The initials

LDSCR

and

HDSCR

mean or refer to low and high dust Selective Catalytic Reduction, respectively.

(18) The initials

LNB

mean or refer to low NO

burners.

(19) The initials

LTS

mean or refer to Long-term Strategy.

(20) The initials

MACT

mean or refer to Maximum Achievable Control Technology.

(21) The initials

mean or refer to megawatts.

(22) The initials

NAAQS

mean or refer to National Ambient Air Quality Standards.

(23) The initials

NEI

mean or refer to National Emissions Inventory.

(24) The initials

NESCAUM

mean or refer to Northeast States for Coordinated Air Use Management.

(25) The initials

mean or refer to National Monument.

(26) The initials

mean or refer to nitrogen oxides.

(27) The initials

mean or refer to National Park.

(28) The initials

NPS

mean or refer to the National Park Service.

(29) The initials

NSCR

mean or refer to non-selective catalytic reduction.

(30) The initials

NSPS

mean or refer to new source performance standards.

(31) The initials

mean or refer to particulate matter.

(32) The initials

2.5

mean or refer to fine particulate matter with an aerodynamic diameter of less than 2.5 micrometers.

(33) The initials

mean or refer to particulate matter with an aerodynamic diameter of less than 10 micrometers.

(34) The initials

PSAT

mean or refer to Particulate Source Apportionment Technology.

(35) The initials

PSD

mean or refer to Prevention of Significant Deterioration.

(36) The initials

PTE

mean or refer to potential to emit.

(37) The initials

mean or refer to regional haze.

(38) The initials

RHR

mean or refer to the Regional Haze Rule, originally promulgated in 1999 and codified at 40 CFR 51.301-309.

(39) The initials

RMC

mean or refer to Regional Modeling Center.

(40) The initials

mean or refer to Reasonable Progress.

(41) The initials

RPG

RPGs

mean or refer to Reasonable Progress Goal(s).

(42) The initials

SCR

mean or refer to Selective Catalytic Reduction.

(43) The initials

SIP

mean or refer to State Implementation Plan.

(44) The initials

SNCR

mean or refer to Selective Non-catalytic Reduction.

(45) The initials

mean or refer to sulfur dioxide.

(46) The initials

SOFA

mean or refer to Separated Overfire Air.

(47) The initials

SRP

mean or refer to Salt River Project Agricultural Improvement and Power District.

(48) The initials

tpy

mean tons per year.

(49) The initials

TSD

mean or refer to Technical Support Document.

(50) The initials

TSF

mean or refer to tons of stone feed.

(51) The initials

ULNB

mean or refer to ultra-low NO

burners.

(52) The initials

URP

mean or refer to Uniform Rate of Progress.

(53) The initials

VOC

mean or refer to volatile organic compounds.

(54) The initials

WRAP

mean or refer to the Western Regional Air Partnership.

B. Docket

This proposed action relies on documents, information and data that are listed in the index on

http://www.regulations.gov

under docket number EPA-R09-OAR-2013-0588. Previous proposed and final actions regarding Arizona's RH SIP are under docket number EPA-R09-OAR-2012-0904 and EPA-R09-OAR-2012-0021. Although listed in the index, some information is not publicly available (e.g., Confidential Business Information (CBI)). Certain other material, such as copyrighted material, is publicly available only in hard copy form. Publicly available docket materials are available either electronically at

http://www.regulations.gov

or in hard copy at the Planning Office of the Air Division, AIR-2, EPA Region 9, 75 Hawthorne Street, San Francisco, CA 94105. EPA requests that you contact the individual listed in the

FOR FURTHER INFORMATION CONTACT

section to view the hard copy of the docket. You may view the hard copy of the docket Monday through Friday, 9-5 PST, excluding Federal holidays.

C. Instructions for Submitting Comments to EPA

Written comments must be submitted on or before March 31, 2014. Submit your comments, identified by Docket ID No. EPA-R09-OAR-2013-0588, by one of the following methods:

•

Federal Rulemaking portal: http://www.regulations.gov

. Follow the on-line instructions for submitting comments.

•

Email: r9azreghaze@epa.gov.

•

Fax:

415-947-3579 (Attention: Thomas Webb).

•

Mail, Hand Delivery or Courier:

Thomas Webb, EPA Region 9, Air Division (AIR-2), 75 Hawthorne Street, San Francisco, California 94105. Hand and courier deliveries are only accepted Monday through Friday, 8:30 a.m. to 4:30 p.m., excluding Federal holidays. Special arrangements should be made for deliveries of boxed information.

EPA's policy is to include all comments received in the public docket without change.

We may make comments available online at

http://www.regulations.gov

, including any personal information provided, unless the comment includes information claimed to be CBI or other information for which disclosure is restricted by statute. Do not submit information that you consider to be CBI or that is otherwise protected through

http://www.regulations.gov

or email. The

http://www.regulations.gov

Web site is an “anonymous access” system, which means EPA will not know your identity or contact information unless you provide it in the body of your comment. If you send an email comment directly to EPA, without going through

http://www.regulations.gov

, we will include your email address as part of the comment that is placed in the public docket and made available on the Internet. If you submit an electronic comment, EPA recommends that you include your name and other contact information in the body of your comment and with any disk or CD-ROM you submit. If EPA cannot read your comment due to technical difficulties and cannot contact you for clarification, EPA may not be able to consider your comment. Electronic files should not include special characters or any form of encryption, and be free of any defects or viruses.

D. Submitting Confidential Business Information

Do not submit CBI to EPA through

http://www.regulations.gov

or email. Clearly mark the part or all of the information that you claim as CBI. For CBI information in a disk or CD ROM that you mail to EPA, mark the outside of the disk or CD ROM as CBI and identify electronically within the disk or CD ROM the specific information that is claimed as CBI. In addition to one complete version of the comment that includes information claimed as CBI, you must submit a copy of the comment that does not contain the information claimed as CBI for inclusion in the public docket. We will not disclose information so marked except in accordance with procedures set forth in 40 CFR part 2.

E. Tips for Preparing Comments

When submitting comments, remember to:

• Identify the rulemaking by docket number and other identifying information (e.g., subject heading,

Federal Register

date and page number).

• Explain why you agree or disagree; suggest alternatives and substitute language for your requested changes.

• Describe any assumptions and provide any technical information and/or data that you used.

• If you estimate potential costs or burdens, explain how you arrived at your estimate in sufficient detail to allow for it to be reproduced.

• Provide specific examples to illustrate your concerns, and suggest alternatives.

• Explain your views as clearly as possible, avoiding profanity or personal threats.

• Make sure to submit your comments by the identified comment period deadline.

To provide opportunities for questions and discussion, EPA will hold an open house prior to the public hearing. During the open house, EPA staff will be available informally to

answer questions on our proposed rule. Any comments made to EPA staff during the open house must still be provided formally in writing or orally during a public hearing to be considered in the record. The open house and public hearing schedule is as follows.

F. Public Hearings

EPA will hold two public hearings at the dates, times and locations stated below to accept oral and written comments into the record. To request interpretation services or to request reasonable accommodation for a disability, please contact the person in the

FOR FURTHER INFORMATION CONTACT

section by February 14, 2014.

Public Hearing in Phoenix:

Date:

February 25, 2014.

Open House:

4-5 p.m.

Public Hearing:

6-8 p.m.

Location:

Phoenix Convention Center, Rooms 150-153, 33 South 3rd Street, Phoenix, Arizona 85004.

Public Hearing in Tucson:

Date:

February 26, 2014.

Open House:

4-5 p.m.

Public Hearing:

6-8 p.m.

Location:

Tucson High Magnet School, Auditorium, 400 North 2nd Avenue, Tucson, Arizona 85705.

The public hearing will provide the public with an opportunity to present views or information concerning the proposed RH FIP for Arizona. EPA may ask clarifying questions during the oral presentations, but will not respond to the presentations at that time. We will consider written statements and supporting information submitted during the comment period with the same weight as any oral comments and supporting information presented at the public hearing. Please consult section I.C, I.D and I.E of this preamble for guidance on how to submit written comments to EPA. We will include verbatim transcripts of the hearing in the docket for this action. The EPA Region 9 Web site for the rulemaking, which includes the proposal and information about the public hearing, is at

http://www.epa.gov/region9/air/actions

II. Proposed Actions Background and Overview

A. Background

The Clean Air Act (CAA) establishes as a national goal the prevention of any future, and the remedying of any existing man-made impairment of visibility in 156 national parks and wilderness areas designated as Class I areas. Arizona has a wealth of such areas. The sources addressed in this FIP affect many Class I areas in the State of Arizona and adjacent states. This FIP will ensure that progress is made toward natural visibility conditions at these national treasures, as Congress intended when it directed EPA to improve visibility in national parks and wilderness areas. Please refer to our previous rulemaking on the Arizona RH SIP for additional background regarding the CAA, regional haze and EPA's RHR.

77 FR 75704, 75707-75702 (December 21, 2012).

B. Regional Haze

We propose to promulgate a FIP as described in this notice and summarized in this section to address those portions of Arizona's RH SIP that we disapproved on July 30, 2013.

We disapproved in part Arizona's BART control analyses and determinations for four sources, Reasonable Progress Goal (RPG) analyses and determinations, and Long-term Strategy (LTS) for making reasonable progress. The proposed FIP includes emission limits, compliance schedules and requirements for equipment maintenance, monitoring, testing, recordkeeping and reporting for all affected sources and units. The regulatory language for the proposed FIP requirements is under Part 52 at the end of this notice.

78 FR 46142.

1. Proposed BART Determinations

EPA conducted BART analyses and determinations for four sources: Sundt Generating Station Unit 4, Nelson Lime Plant Kilns 1 and 2, Hayden Smelter and Miami Smelter. The results of our BART evaluations are summarized here for each source and are shown in Table 1. We are seeking comments on our proposals.

Sundt:

We propose that Sundt Unit 4 is BART-eligible and subject to BART for sulfur dioxide (SO

), nitrogen oxides (NO

) and particulate matter with aerodynamic diameter less than 10 micrometers (PM

). For NO

, we propose an emission limit of 0.36 lb/MMBtu as BART based upon an annual capacity factor of 0.49, which is consistent with the use of Selective Non-Catalytic Reduction (SNCR) as a control technology. For SO

, we propose an emission limit of 0.23 lb/MMBtu as BART on a 30-day boiler operating day (BOD) rolling basis, which is consistent with dry sorbent injection (DSI) as a control technology. For PM

, we propose a filterable PM

emission limit of 0.030 lb/MMBtu as BART based on the use of the existing fabric filter baghouse. We also are proposing a switch to natural gas as a better-than-BART alternative to the other proposed controls for all three pollutants.

Nelson Lime Plant:

We propose that Nelson Lime Kilns 1 and 2 are subject to BART for NO

and PM

. For NO

, we propose a BART emission limit at Kiln 1 of 3.80 lb/ton lime and at Kiln 2 of 2.61 lb/ton lime on a 30-day rolling basis as verified by continuous emission monitoring systems (CEMS). This emission limit is consistent with the use of low-NO

burners (LNB) and SNCR as control technologies. We propose that BART for SO

is an emission limit of 9.32 lb/ton for Kiln 1 and 9.73 lb/ton for Kiln 2 on a 30-day rolling basis, which is consistent with the use of a lower sulfur fuel blend. For PM

, we propose a BART emission limit of 0.12 lb/tons of stone feed (TSF) to control PM

at Kilns 1 and 2 based on the use of the existing fabric filter baghouses. This level of control is commensurate with the MACT standard that applies to this source.

Hayden Smelter:

We propose that the Hayden Smelter is subject to BART for NO

, and propose BART emission limits for NO

and SO

. EPA previously approved the State's determination that the Hayden Smelter is subject to BART for SO

. For NO

, we propose to find that controlling emissions from the converters and anode furnaces is cost-effective, but would not result in sufficient visibility improvement to warrant the cost. Therefore, we are proposing an annual emission limit of 40 tpy NO

emissions from the BART-eligible units, which is consistent with current emissions from these units. For SO

from the converters, we propose a BART control efficiency of 99.8 percent on a 30-day rolling basis on all SO

captured by primary and secondary control systems, which can be achieved with a new double contact acid plant. For SO

from the anode furnaces, we propose to find that controlling the 37 tons per year (tpy) of SO

emissions from these furnaces, while cost-effective, is not warranted as BART given the potential for only minimal visibility improvement. We propose as an emission limitation for the anode furnace a work practice standard requiring that the furnaces only be charged with blister copper or higher purity copper. We previously approved Arizona's determination that BART for PM

at the Hayden Smelter is no additional controls. In order to ensure the enforceability of this determination, we are proposing to incorporate emission limitations and associated compliance requirements from the National Emission Standard for Hazardous Air Pollutants (NESHAP) for

Primary Copper Smelting at 40 CFR Part 63, Subpart QQQ, as part of the LTS.

Miami Smelter:

EPA proposes that the Miami Smelter is subject to BART for NO

, and proposes BART emission limits for NO

and SO

. EPA previously approved the State's determination that the Miami Smelter is subject to BART for SO

. For NO

, we propose to find that controlling the small amount of emissions from the converters and electric furnace is cost-effective, but would not result in sufficient visibility improvement to warrant the cost. Therefore, we are proposing an annual emission limit of 40 tpy NO

emissions from the BART-eligible units, which is consistent with current emissions. For SO

from the converters, we propose a BART control efficiency of 99.7 percent on a 30-day rolling basis on all SO

emissions captured by the primary and secondary control systems as verified by CEMS. This control efficiency could be met through improvements to the primary capture system, construction of a secondary capture system, and application of the MACT QQQ standards to the capture systems. For SO

emissions from the electric furnace, we propose as BART the work practice standard to prohibit active aeration. We previously approved Arizona's determination that BART for PM

at the Miami Smelter is the NESHAP for Primary Copper Smelting. We now propose to find that the federally enforceable provisions of the NESHAP, which apply to the Miami Smelter and are incorporated into its Title V Permit, are sufficient to ensure the enforceability of this determination.

Table 1—Proposed Emission Limits on BART Sources

Source

Units

Pollutants

Limit

Measure

Corresponding control technology

Sundt Generating Station

Unit 4

0.36

0.23

0.030

lb/MMBtu

Selective Non-Catalytic Reduction.

Dry Sorbent Injection.

Fabric filter baghouse (existing).

Unit 4 (Alternative)

0.25

0.00064

0.010

lb/MMBtu

Switch to natural gas.

Chemical Lime Nelson

Kiln 1

3.80

9.32

0.12

lb/ton feed

Selective Non-Catalytic Reduction.

Lower sulfur fuel.

Fabric filter baghouse (existing).

Kiln 2

2.61

9.73

0.12

Selective Non-Catalytic Reduction.

Lower sulfur fuel.

Fabric filter baghouse (existing).

Hayden Smelter

Converters 1, 3-5

99.8

tpy

Control efficiency

None.

New double contact acid plant.

Anode Furnaces 1, 2

None

Work practice standard.

Miami Smelter

Converters 2-5

99.7

tpy

Control efficiency

None.

Improve primary and new secondary capture systems.

Electric Furnace

None

Work practice standard.

2. Proposed RP Determinations

Point Sources of NO

: EPA conducted an extensive RP analysis of NO

point sources that resulted in proposed determinations for nine sources and proposed controls on two sources as shown in Table 2. We are proposing an emissions limit of 2.12 lb/ton on Kiln 4 of the Phoenix Cement Clarkdale Plant based on a 30-day rolling average, which is consistent with SNCR as a control technology. We are proposing an emissions limit of 2.67 lb/ton on Kiln 4 of the CalPortland Cement Rillito Plant based on a 30-day rolling average, which also is consistent with SNCR control technology. We are also taking comment on the possibility of requiring a rolling 12-month cap on NO

emissions in lieu of a lb/ton emission limit. For Phoenix Cement, this cap would be 947 tpy and apply to Kiln 4. For CalPortland, this cap would be 2,082 tpy and apply to Kilns 1-4.

Table 2—Proposed Emission Limits on RP Sources

Source

Units

Pollutants

Limit

Measure

Corresponding control technology

Phoenix Cement

Kiln 4

2.12

lb/ton

Selective Non-Catalytic Reduction.

CalPortland Cement

Kiln 4

2.67

lb/ton

Selective Non-Catalytic Reduction.

Area Sources of NO

and SO

: We propose to find that it is reasonable not to require additional controls on these sources at this time. Primarily, these area source categories are distillate fuel oil combustion in industrial and commercial boilers and in internal combustion engines, and residential natural gas combustion. The State's area sources, which currently contribute a relatively small percentage of the visibility impairment at impacted Class I areas, would benefit from better emission inventories and an improved RP analysis in the next planning period.

Reasonable Progress Goals:

EPA is proposing RPGs consistent with a combination of control measures that include those in the approved Arizona RH SIP as well as the approved and proposed Arizona RH FIP. While not quantifying a new set of RPGs based on these control measures, we propose that it is reasonable to assume improved levels of visibility at Arizona's 12 Class I areas by 2018 since the measures in the FIP are significantly beyond what was in the State's plan.

Demonstration of Reasonable Progress:

EPA proposes to find that it is not reasonable to provide for rates of progress at the 12 Class I areas consistent with the uniform rate of progress (URP) in this planning period.

Given the variety and location of sources contributing to visibility impairment in Arizona, EPA considers

it unlikely that Arizona's Class I areas will meet the URP in 2018. We propose to find that the RP analyses underlying our actions on the Arizona SIP

and in this proposal are sufficient to demonstrate that it is not reasonable to provide for rates of progress in this planning period that would attain natural conditions by 2064.

This is consistent with our proposed and final rules on the Arizona RH SIP in which we approved Arizona's determinations that it is not reasonable to require additional controls to address organic carbon, elemental carbon, coarse mass and fine soil during this planning period.

We also approved the State's decision not to require additional controls (i.e., controls beyond what the State or we determine to be BART) on point sources of SO

40 CFR 51.308(d)(1)(ii).

See proposed actions at 77 FR 75727-75730, 78 FR 29297-292300 and final action at 78 FR 46172.

40 CFR 51.308(d)(1)(ii).

See 77 FR 75728 for a discussion on sources of organic carbon and elemental carbon (fires), and 78 FR 29297-29299 for a discussion of coarse mass and fine soil.

78 FR 46172.

3. Long-Term Strategy Proposal

EPA proposes to find that provisions in today's proposal in combination with provisions in the approved Arizona SIP and FIP

fulfill the requirements of 40 CFR 51.308(d)(3)(ii), (v)(C) and (v)(F). These requirements are to include in the LTS measures needed to achieve emission reductions for out-of-state Class I areas, emissions limitations and schedules for compliance to achieve the reasonable progress goals, and enforceability of emissions limitations and control measures.

In today's notice we propose to promulgate emission limits, compliance schedules and other requirements for four BART sources and two RP sources to complete the actions taken in our previous final rule to address these requirements.

77 FR 75512-72580, December 5, 2012.

See 78 FR 46173 (codified at 40 CFR 52.145(e)(ii)).

C. Interstate Transport of Pollutants That Affect Visibility

We propose that a combination of SIP and FIP measures will satisfy the FIP obligation for the visibility requirement of CAA section 110(a)(2)(D)(i)(II) for the 1997 8-hour ozone, 1997 PM

2.5

, and 2006 PM

2.5

NAAQS. CAA section 110(a)(2)(D)(i)(II) requires that all SIPs contain adequate provisions to prohibit emissions that will interfere with other states' required measures to protect visibility. We refer to this requirement herein as the interstate transport visibility requirement. ADEQ submitted SIP revisions to address this requirement in 2007 for the 1997 8-hour ozone NAAQS

and 1997 PM

2.5

NAAQS

(2007 Transport SIP)

and in 2009 for the 2006 PM

2.5

NAAQS

(2009 Transport SIP).

Each of these SIP revisions indicated that it is appropriate to assess Arizona's interference with other states' measures to protect visibility in conjunction with the State's RH SIP.

In our final rule published on July 30, 2013, EPA disapproved these SIP submittals with respect to the interstate transport visibility requirement, triggering the obligation for EPA to promulgate a FIP to address this requirement.

Accordingly, today's notice describes our proposed FIP for the interstate transport visibility requirement for the 1997 8-hour ozone, 1997 PM

2.5

, and 2006 PM

2.5

NAAQS.

62 FR 38856, July 18, 1997.

62 FR 38652, July 18, 1997.

“Revision to the Arizona State Implementation Plan Under Clean Air Act Section 110(a)(2)(D)(i)—Regional Transport,” submitted by ADEQ on May 24, 2007.

71 FR 61144, October 17, 2006.

“Arizona State Implementation Plan Revision under Clean Air Act Section 110(a)(1) and (2); 2006 PM

2.5

NAAQS, 1997 PM

2.5

NAAQS, and 1997 8-hour Ozone NAAQS,” submitted by ADEQ on October 14, 2009, which addressed the requirements of section 110(a)(2)(D)(i) with respect to the 2006 PM

2.5

NAAQS in Section 2.4 and Appendix B of the submittal.

This concept is also presented in EPA's 2006 guidance memo on interstate transport, which recommended that states make a submission indicating that it was premature, at that time, to determine whether there would be any interference with other states' required measures to protect visibility until the submission and approval of regional haze SIPs. See “Guidance for State Implementation Plan (SIP) Submissions to Meet Current Outstanding Obligations Under Section 110(a)(2)(D)(i) for the [1997] 8-Hour Ozone and PM

2.5

National Ambient Air Quality Standards,” August 15, 2006.

78 FR 46142, July 30, 2013.

III. Review of State and EPA Actions on Regional Haze

A. EPA's Schedule To Act on Arizona's RH SIP

EPA received a notice of intent to sue in January 2011 stating that we had not met the statutory deadline for promulgating RH FIPs and/or approving RH SIPs for dozens of states, including Arizona. This notice was followed by a lawsuit filed by several advocacy groups (Plaintiffs) in August 2011.

In order to resolve this lawsuit and avoid litigation, EPA entered into a Consent Decree with the Plaintiffs, which sets deadlines for action for all of the states covered by the lawsuit, including Arizona. This decree was entered and later amended by the United States District Court for the District of Columbia over the opposition of Arizona.

Under the terms of the Consent Decree, as amended, EPA is currently subject to three sets of deadlines for taking action on Arizona's RH SIP as listed in Table 3.

National Parks Conservation Association

Jackson

(D.D.C. Case 1:11-cv-01548).

National Parks Conservation Association

Jackson

(D.D.C. Case 1:11-cv-01548), Memorandum Order and Opinion (May 25, 2012), Minute Order (July 2, 2012), Minute Order (November 13, 2012) and Minute Order (February 15, 2013).

Id.

Table 3—Consent Decree Deadlines for EPA To Act on Arizona's RH SIP

EPA actions

Proposed rule

Final rule

Phase 1—BART determinations for Apache, Cholla and Coronado

July 2, 2012

November 15, 2012.

Phase 2—All remaining elements of the Arizona RH SIP

December 8, 2012

July 15, 2013.

Phase 3—FIP for disapproved elements of the Arizona RH SIP

January 27, 2014

June 27, 2014.

Published in the

Federal Register

on July 20, 2012, 77 FR 42834.

Published in the

Federal Register

on December 5, 2012, 77 FR 72512.

Published in the

Federal Register

on December 21, 2012, 77 FR 75704.

Published in the

Federal Register

on July 30, 2013, 78 FR 46142.

B. History of State Submittals and EPA Actions

Because four of Arizona's 12 mandatory Class I Federal areas are on the Colorado Plateau, the State had the option of submitting a RH SIP under CAA section 309 of the RHR. A SIP that is approved by EPA as meeting all of the requirements of section 309 is “deemed to comply with the requirements for reasonable progress with respect to the 16 Class I areas [on the Colorado

Plateau] for the period from approval of the plan through 2018.”

When these regulations were first promulgated, 309 SIPs were due no later than December 31, 2003. Accordingly, ADEQ submitted to EPA on December 23, 2003, a 309 SIP for Arizona's four Class I Areas on the Colorado Plateau. ADEQ submitted a revision to its 309 SIP, consisting of rules on emissions trading and smoke management, and a correction to the State's regional haze statutes, on December 31, 2004. EPA approved the smoke management rules submitted as part of the revisions in 2004,

but did not propose or take final action on any other portion of the 309 SIP.

40 CFR 51.309(a).

71 FR 28270 and 72 FR 25973.

In response to a court decision,

EPA revised 40 CFR 51.309 on October 13, 2006, making a number of substantive changes and requiring states to submit revised 309 SIPs by December 17, 2007.

Subsequently, ADEQ sent a letter to EPA dated December 24, 2008, acknowledging that it had not submitted a SIP revision to address the requirements of 40 CFR 51.309(d)(4) related to stationary sources and 40 CFR 51.309(g), which governs reasonable progress requirements for Arizona's eight mandatory Class I areas outside of the Colorado Plateau.

Center for Energy and Economic Development

EPA,

398 F.3d 653 (D.C. Circuit 2005).

71 FR 60612.

Letter from Stephen A. Owens, ADEQ, to Wayne Nastri, EPA, dated December 24, 2008.

EPA made a finding on January 15, 2009, that 37 states, including Arizona, had failed to make all or part of the required SIP submissions to address regional haze.

Specifically, EPA found that Arizona failed to submit the plan elements required by 40 CFR 51.309(d)(4) and (g). EPA sent a letter to ADEQ on January 14, 2009, notifying the State of this failure to submit a complete SIP. ADEQ decided to submit a SIP under CAA section 308, instead of under section 309. EPA proposed on February 5, 2013,

to disapprove Arizona's 309 SIP except for the smoke management rules that we had previously approved. Our final rule partially disapproving Arizona's 309 SIP was published on August 8, 2013.

74 FR 2392.

78 FR 8083.

78 FR 48326.

ADEQ adopted and transmitted its 2011 RH SIP under section 308 of the RHR to EPA Region 9 in a letter dated February 28, 2011. The SIP was determined complete by operation of law on August 28, 2011.

The SIP was properly noticed by the State and available for public comment for 30 days prior to one public hearing held in Phoenix, Arizona, on December 2, 2010. Arizona included in its SIP responses to written comments from EPA Region 9, the National Park Service, the U.S. Forest Service, and other stakeholders including regulated industries and environmental organizations. The 2011 RH SIP is available to review in the docket for this proposed rule.

CAA section 110(k)(1)(B).

“Arizona State Implementation Plan, Regional Haze under Section 308 of the Federal Regional Haze Rule,” February 28, 2011.

As shown in Table 3, the first phase of EPA's action on the 2011 RH SIP addressed three BART sources. The final rule for the first phase (a partial approval and partial disapproval of the State's plan and a partial FIP) was signed by the Administrator on November 15, 2012, and published in the

Federal Register

on December 5, 2012. The emission limits on the three sources will improve visibility by reducing NO

emissions by about 22,700 tons per year. In the second phase of our action, we proposed on December 21, 2012, to approve in part and disapprove in part the remainder of the 2011 RH SIP. ADEQ submitted an Arizona RH SIP Supplement on May 3, 2013, to correct certain deficiencies identified in that proposal. We then proposed on May 20, 2013, to approve in part and disapprove in part the Supplement. Our final rule approving in part and disapproving in part Arizona's RH SIP was published on July 30, 2013.

C. EPA's Authority To Promulgate a FIP

Under CAA section 110(c), EPA is required to promulgate a FIP within 2 years of the effective date of a finding that a state has failed to make a required SIP submission. The FIP requirement is terminated if a state submits a regional haze SIP, and EPA approves that SIP before promulgating a FIP. See 74 FR 2392. Specifically, CAA section 110(c) provides:

(1) The Administrator shall promulgate a Federal implementation plan at any time within 2 years after the Administrator—

(A) finds that a State has failed to make a required submission or finds that the plan or plan revision submitted by the State does not satisfy the minimum criteria established under [CAA section 110(k)(1)(A)], or

(B) disapproves a State implementation plan submission in whole or in part, unless the State corrects the deficiency, and the Administrator approves the plan or plan revision, before the Administrator promulgates such Federal implementation plan.

Section 302(y) defines the term “Federal implementation plan” in pertinent part, as:

[A] plan (or portion thereof) promulgated by the Administrator to fill all or a portion of a gap or otherwise correct all or a portion of an inadequacy in a State implementation plan, and which includes enforceable emission limitations or other control measures, means or techniques (including economic incentives, such as marketable permits or auctions or emissions allowances) . . .

Thus, because we determined that Arizona failed to timely submit a Regional Haze SIP, we are required to promulgate a Regional Haze FIP for Arizona, unless we first approve a SIP that corrects the non-submittal deficiencies identified in our finding of January 15, 2009. For the reasons explained below, we approved in part and disapproved in part the Arizona Regional Haze SIP on July 30, 2013. Therefore, we are proposing a FIP to address those portions of the SIP that we disapproved.

IV. EPA's BART Process

A. BART Factors

The purpose of the BART analysis is to identify and evaluate the best system of continuous emission reduction based on the BART Guidelines

as summarized below. Steps 1 through 3 address the availability, feasibility and effectiveness of retrofit control options. In our analysis of control technology options, we expressly include the emission baseline calculation that is a key factor in determining control effectiveness. Step 4 is the five-factor BART analysis that results in selecting the emission limit that represents BART in Step 5. Following the process steps is a short description of each BART factor.

See July 6, 2005 BART Guidelines, 40 CFR 51, Regional Haze Regulations and Guidelines for Best Available Retrofit Technology Determinations.

Step 1—Identify all available retrofit control technologies.

Step 2—Eliminate technically infeasible options.

Step 3—Evaluate control effectiveness of remaining control technologies.

Step 4—Evaluate impacts and document the results.

• Factor 1: Cost of compliance.

• Factor 2: Energy and non-air quality environmental impacts of compliance.

• Factor 3: Pollution control equipment in use at the source.

• Factor 4: Remaining useful life of the facility.

• Factor 5: Visibility impacts.

Step 5—Select BART.

Factor 1: Costs of Compliance:

The evaluation of costs is an important part of a five-factor analysis because it influences the cost-effectiveness that is

compared to the visibility benefits. Estimating the cost of compliance primarily depends on the cost estimates and control effectiveness of each technically feasible BART control option. For each of the four BART facilities evaluated in this section, we state the source of the cost-related information and how it was used in our analysis. While EPA relies primarily on the cost methods in our Control Cost Manual, we also rely on verified cost estimates from the companies and cost methods used for specific industries. In some cases, certain capital costs and annual operating costs were developed by our contractor based on actual costs associated with specific types of sources. Where possible, we have conducted new cost analyses considering more recent information from ADEQ or from the four BART facilities. Please refer to the TSD for the detailed cost analyses.

Factor 2: Energy and Non-air Quality Environmental Impacts:

In assessing the potential energy impacts of BART control options, we consider direct and indirect effects on energy availability and costs. An example of a direct energy impact is the cost of energy consumption from the control equipment. Examples of non-air quality impacts include safety issues associated with handling and transportation of anhydrous ammonia or the ability to sell fly ash rather than dispose of it.

Factor 3: Pollution Equipment in Use at the Source:

The presence of existing pollution control technology at each source is reflected in our BART analysis in two ways. First, we always consider simple retention of existing equipment as a BART candidate. We also consider existing equipment in determining available control technologies that can be used with or replace such equipment. Second, where appropriate, we consider existing equipment in developing baseline emission rates for use in cost calculations and visibility modeling. Pollutant-specific discussions of these issues are included in the following sections.

Factor 4: Remaining Useful Life of the Source:

We consider each source's “remaining useful life” as one element of the overall cost analysis as allowed by the BART Guidelines.

In cases where we are not aware of any enforceable shut-down date for a particular source or unit, we use a 20-year amortization period as the remaining useful life per the EPA Cost Control Manual.

40 CFR Part 51, Appendix Y, section IV.D.4.k.

Factor 5: Anticipated Degree of Visibility Improvement:

EPA relied on the CALPUFF modeling system (version 5.8) for visibility modeling, which consists of the CALPUFF dispersion model, the CALMET meteorological data processor, and the CALPOST post-processing program. The initial modeling was performed by our contractor, the University of North Carolina (UNC) at Chapel Hill. In some cases, companies submitted BART analyses including visibility modeling that we used to evaluate visibility benefits. An explanation of the visibility analysis and tables follows this section, a description of the modeling is included in the five-factor discussion for each source, and more details are available in the TSD.

B. Visibility Analysis

EPA estimated the degree of visibility improvement expected to result from various BART control options based on the difference between baseline visibility impacts prior to controls and visibility impacts with controls in operation. Baseline emissions were based on the highest 24-hour emissions from monitored emissions data when available, otherwise from estimates of production rates and emission factors. Control case emissions were derived from the baseline by applying the percent reduction in emission factor expected from the control. Impacts at all Class I areas within 300 km of each facility were assessed. EPA used the CALPUFF model version 5.8

to determine the baseline and post-control visibility impacts, following the modeling approach recommended in the BART Guidelines. Our contractor at UNC developed a modeling protocol and carried out most of the modeling and the post-processing of model output into tables of visibility impacts. EPA supplemented this for certain sources with modeling of additional control scenarios, corrections to some scenarios and post-processing work, and some sensitivity simulations. Also, EPA performed the modeling for the two smelters. Details of the modeling are in the TSD.

EPA relied on version 5.8 of CALPUFF because it is the EPA-approved version promulgated in the Guideline on Air Quality Models (40 CFR part 51, Appendix W, section 6.2.1.e; 68 FR 18440, April 15, 2003). EPA updated the specific version to be used for regulatory purposes on June 29, 2007, including minor revisions as of that date; the approved CALPUFF modeling system includes CALPUFF version 5.8, level 070623, and CALMET version 5.8 level 070623. At this time, any other version of the CALPUFF modeling system would be considered an “alternative model”, subject to the provisions of Guideline on Air Quality Models section 3.2.2(b), requiring a full theoretical and performance evaluation.

EPA modeled all units (stacks) and pollutants simultaneously for each source. Modeling of all emissions from all units accounts for the chemical interaction between multiple plumes, and between plumes and background concentrations. This also accounts for the fact that deciview benefits from controls on individual units are not strictly additive. As recommended in the BART Guidelines, the 98th percentile daily impact in deciviews is used as the basic metric of visibility impact. EPA relied on the 98th percentile over the merged 2001-2003 period. The alternative of using the average of the three 98th percentiles from 2001, 2002 and 2003 was also calculated, and the results of using it are provided in the TSD, although they differ little from the merged approach. Both are valid indicators of the 98th percentile.

EPA also mainly relied on the revised IMPROVE equation for translating pollutant concentrations into deciviews (CALPOST visibility method 8), the recommended method for new visibility analyses. The old IMPROVE equation (method 6) was used by most states in their original SIP submittals and was acceptable at that time. EPA used the best 20 percent of natural background days in calculating delta deciviews. For the original SIP submittals, states were free to use this or the annual average background. Overall, we refer to the method we used as method “8b” (“b” for “best”). Model results using visibility method 6 and annual average background conditions (“a” for average) also are provided in the TSD (i.e., methods 6a, 6b, and 8a, as well as 8b).

For each modeled day, the CALPUFF model provides the highest impact from among the receptor locations for a given Class I area. The baseline impact in the tables is the 98th percentile among these daily values. The improvement in the tables is the difference between that baseline impact and the 98th percentile impact after applying controls. The 98th percentile is represented by the 22nd high over the 2001-2003 period modeled. The TSD includes an alternative, the average of each of the three years' 8th highs, which yields slightly different values.

C. Explanation of Visibility Tables

For each facility, this notice provides one or more tables of visibility impacts and visibility improvement from controls in deciviews. Each table has the same format: columns list the Class I areas within 300 km of the facility, the distance,

baseline modeled visibility impact from the facility for each area, and one or more columns with the

modeled visibility improvement from a candidate control option. A modeling run abbreviation, such as “base” or “ctrl2”, is included along with a short description of the option. For several facilities, there are two different baselines incorporating different emission assumptions. For these, there are baseline and control columns for each of the two baselines. For Sundt Unit 4, there are separate tables for SO

and NO

controls, and an additional table showing the effect of reductions for both SO

and NO

for the proposed BART controls and for a better-than-BART alternative. At the bottom of each table are five rows showing impacts and improvements from the facility for all the Class I areas considered together, and also two measures of visibility cost-effectiveness. The cost-effectiveness here is “dollars per deciview,” where dollars is the annualized total cost of the control in millions of dollars per year, divided by either the sum of deciview improvements over all impacted Class I areas, or the largest single area deciview improvement. Cost-effectiveness in terms of dollars per ton is presented in other tables and has been considered for each source and BART option. The headings for these table rows are:

The distances given are from the facility to the nearest model receptor location; distances to the actual Class I area boundary may be slightly less. Receptor locations are defined for all Class I areas by the National Park Service. See “Class I Receptors” Web site,

http://www2.nature.nps.gov/air/maps/Receptors/

(1) “Cumulative (sum),” the cumulative impact or improvement that is computed as the sum of impact or improvement over all the areas;

(2) “Maximum,” single largest impact or improvement that is the maximum over all the areas;

(3) “# CIAs >= 0.5 dv,” the number of Class I areas having a baseline impact from the source of at least 0.5 dv (or, for the control columns, the number of areas showing improvement of at least 0.5 dv due to the control);

(4) “Million $/dv (cumul. dv),” annual control cost in millions of dollars per deciview considering the improvement at all the Class I areas together; and

(5) “Million $/dv (max. dv),” annualized cost per deciview considering the largest single area improvement.

The Federal Land Managers have sometimes used $10 million/dv as a comparison benchmark for the $/dv computed from the maximum, and $20 million/dv as a benchmark for $/dv computed from cumulative deciviews. We have not endorsed the use of these or any other $/dv benchmarks as criteria for making BART determinations.

The TSD for this notice provides bar charts and additional visibility tables, including results for individual modeled years and their average, the old IMPROVE equation, and annual average background conditions instead of best 20 percent. There also are model results for various sensitivity analyses.

V. EPA's Proposed BART FIP

A. Sundt Generating Station Unit 4

Summary:

EPA is proposing to find that Sundt Unit 4 is eligible for and subject to BART. EPA is proposing BART emissions limits on Sundt Generating Station Unit 4 for NO

and PM

based on the corresponding control technologies listed in Table 4 and described in the following BART analyses. For NO

, we propose an emission limit of 0.36 lb/MMBtu consistent with the use of SNCR. For SO

, we propose an emission limit of 0.23 lb/MMBtu consistent with the use of DSI. For PM

, we propose a filterable PM

emission limit of 0.03 lb/MMBtu based on the use of the existing fabric filter baghouse. Finally, we are also proposing a switch to natural gas as a better-than-BART alternative.

Table 4—Sundt 4: Summary of Proposed BART Determinations

Pollutant

Emission limit

(lb/MMBtu)

Control technology

0.36

Selective Non-Catalytic Reduction.

0.23

Dry Sorbent Injection.

0.030

Fabric filter baghouse (existing).

Affected Class I Areas:

Ten Class I areas are within 300 km of Sundt. Their nearest borders range from 17 km to 247 km away, with Saguaro NP the closest, and Galiuro WA the second closest. The highest baseline visibility impact of Sundt Unit 4 is 3.4 dv at Saguaro. The second highest baseline impact is 1.1 dv at Galiuro. Other areas have visibility impacts of 0.5 dv or less. The cumulative sum of visibility impacts over all the Class I areas is 6.6 dv.

Facility Overview:

The Sundt Generating Station is an electric utility power plant located in Tucson, Arizona, operated by Tucson Electric Power. The plant consists of four steam electric boilers and three stationary combustion turbines for a total net generating capacity of approximately 500 megawatts (MW).

Sundt Unit 4 is a steam electric boiler that was manufactured in 1964 and placed into operation in about 1967. Unit 4 is a dry bottom wall-fired boiler with a maximum gross capacity of 130 MW when firing coal. Originally designed to fire natural gas and fuel oil, Sundt Unit 4 was converted to also be able to fire coal in the early 1980s as a result of an order issued by the Department of Energy. The unit now fires both coal and natural gas, as explained in more detail below. As part of the coal conversion, the unit was equipped with a fabric filter for particulate matter control. Unit 4 was upgraded in 1999 with LNB and overfire air (OFA) designed to meet Phase II Acid Rain Program requirements. At present, Unit 4 operates with the pollution control equipment and is subject to the emission limits listed in Table 5 that reflects a coal-operating scenario.

As described in Pima DEQ Permit No. 1052, in the TSD.

Table 5—Sundt 4: Current Emission Limits and Control Technology

Pollutant

Emission limit

Control device

0.46 lb/MMBtu

LNB with OFA.

1 lb/MMBtu

None.

233 lb/hr

Fabric filter/baghouse.

TEP has

indicated that the generating capacity of Sundt Unit 4 while firing coal is reduced compared to its capacity using natural gas. As reported to the Energy Information Agency (EIA), Unit 4 has a 173 MW nameplate capacity while firing natural gas. However, the maximum gross capacity at which the unit could operate for a sustained period of time while burning coal is about 130 MW. This is due primarily to the fact that the amount of coal that can be introduced to the boiler is limited by the size of the boiler. Excess coal injection causes the flame to impinge on the back wall of the boiler which damages the boiler tubes.

A summary of historical emissions data for a recent period of time is in Table 6.

Pima DEQ Permit No. 1052, Attachment F: Phase II Acid Rain Permit.

Pima DEQ Permit No. 1052, Specific Condition II.A.2.b.

As determined by Pima DEQ Permit No. 1052, Specific Condition II.A.1.

TEP's letter dated May 10, 2013, page 2.

Table 6—Sundt 4: Historical Emissions (2008-2012)

Year

Heat duty

(MMBtu/yr)

(tpy)

(lb/MMBtu)

(tpy)

(lb/MMBtu)

Coal

(tons)

Natural gas

(MCF)

2012

6,313,719

945

0.297

371

0.118

44,049

4,660,701

2011

5,993,769

1,366

0.445

2,185

0.729

265,111

157,919

2010

6,869,999

1,303

0.368

1,733

0.505

162,212

1,904,433

2009

4,801,971

709

0.285

636

0.265

73,464

2,642,992

2008

8,709,923

1,880

0.429

2,882

0.661

378,956

18,422

Baseline Emissions Calculations:

The baseline period, baseline emissions, and capacity factor are three key variables in determining BART that are linked to fuel usage. TEP has indicated that while Sundt Unit 4 predominantly has operated as a coal-fired unit, it has recently expanded its use of natural gas as a result of historically low natural gas prices.

As shown in the last column of Table 6, Unit 4 has used much higher amounts of natural gas during 2009-2010 and again in 2012 that are not representative of anticipatable operations based on coal. Accordingly, we use calendar year 2011 emissions when Unit 4 predominately used coal as the baseline period for annual average emission estimates. Although this represents only a single year of emissions data, we consider this period of coal usage, rather than a period of primarily natural gas usage, to represent a realistic depiction of anticipated annual emissions when burning coal.

In addition, we rely on an annual capacity factor of 0.49 based on a coal-fired capacity of 130 MW and actual generation from the baseline period of 2011. For visibility modeling, we used baseline emissions for NO

and SO

based on maximum daily emission rates, as reported to EPA's CAMD Acid Rain Program database, for the period from 2008 to 2010. While this time period is prior to the 2011 baseline period used for the annual emission estimates, the highest daily emission rates from 2008 to 2010 correspond to coal usage. Since these maximum daily emission rates still correspond to coal usage, we consider them reasonable estimates of baseline emissions despite the fact that they are drawn from a baseline period different from the one used to estimate annual emission rates. For PM

10,

the baseline emission rate used in visibility modeling is based on the value in the original Western Regional Air Partnership (WRAP) visibility modeling that reflects the use of coal and the existing fabric filter. For a more detailed analysis of how we determined the baseline period, baseline emissions and capacity factor, please refer to the TSD.

TEP's letter dated May 10, 2013, page 2.

As discussed in the BART Guidelines, 40 CFR Part 51, Appendix Y, section IV.D.4.d.

Modeling Overview:

EPA's contactor UNC performed the initial modeling of Sundt's visibility impacts. EPA performed supplemental modeling to correct some minor errors in the initial work and to estimate impacts from additional control scenarios, such as switching entirely to natural gas fuel. EPA also modeled the impacts for the western unit of Saguaro NP, whereas originally only the eastern unit was included. Although only Unit 4 is BART-eligible, all four Sundt units were included in the CALPUFF modeling to more accurately represent the chemistry of the facility's pollutant plume. Baseline emissions for modeling were based on daily CAMD emissions monitoring data for 2008-2010, a period with no changes in pollution controls at the facility. Control case emissions were derived from the baseline by applying the percent reduction expected from the control.

Saguaro NP has an eastern unit, the Rincon Mountain District, and a western unit, the Tucson Mountain District. In the original set of modeling receptor locations developed by the National Park Service, only the eastern unit was included. CALPUFF modeling typically covered only the eastern unit. This is true of modeling by the WRAP, and also of modeling by EPA's contractor UNC, which used the WRAP work as a starting point. A more recent set of NPS modeling receptors from 2008 is available that covers both eastern and western units of Saguaro. For this FIP, EPA remodeled for both Saguaro units where needed for a given facility. The only facilities for which it makes a significant difference are TEP Sundt and CalPortland Cement due to their close proximity to Saguaro.

1. Proposed Eligible and Subject to BART

EPA is proposing to find that Sundt Unit 4 is eligible for and subject to BART. In our final rulemaking on the Arizona RH SIP dated July 30, 2013, we disapproved ADEQ's finding that Sundt Unit 4 was not eligible for BART.

In particular, we found that, although this unit was “reconstructed” in 1987, it remains BART-eligible because it did not undergo prevention of significant deterioration (PSD) review at the time of reconstruction.

For this reason, we propose to find Sundt Unit 4 is eligible for a BART analysis of the three haze-causing pollutants: NO

, SO

and PM

78 FR 46175 (codified at 40 CFR 52.145(e)(2)(i)).

See 78 FR 75722, 78 FR 46151, and “TEP Sundt Unit I4 BART Eligibility Memo” (November 21, 2012).

Under the RHR and the BART Guidelines, any BART-eligible source that either “causes” or “contributes” to visibility impairment at any Class I area is subject to BART.

EPA previously approved ADEQ's decision to set 0.5 dv as the threshold for determining whether a source contributes to visibility impairment at a given Class I area.

In order to determine whether Sundt Unit 4 is subject to BART, EPA's contractor UNC evaluated whether Unit 4 has an impact of 0.5 dv or more at any Class I area. UNC's visibility modeling showed that two Class I areas experienced a 98th percentile impact greater than 0.5 dv due to emissions from Sundt Unit 4.

In particular, the 98th percentile impact across the three years modeled was 2.798 dv at Saguaro and 0.839 dv at Galiuro.

These results indicate that Sundt Unit 4 causes visibility impairment at Saguaro and contributes to impairment at Galiuro. Therefore, EPA proposes to find that Sundt Unit 4 is subject to BART.

40 CFR part 51, appendix Y, section III.A.

77 FR 46152-53.

Technical Analysis for Arizona and Hawaii Regional Haze FIPs: Report on Identification of Sources Subject to BART, UNC, July 20, 2012, Table 4.

For an expanded discussion of our approach to visibility modeling, please refer to Section III (General Approach to the Five-Factor BART analysis) of the Sundt4 TSD. This approach was used in both determining whether Sundt 4 was subject to BART, as well as in evaluating the visibility factor in the BART analysis.

2. Proposed BART Analysis and Determination for NO

For our NO

BART analysis, we identify all available control technologies, eliminate options that are not technically feasible, and evaluate the control effectiveness of the remaining control options. We then evaluate each technically feasible control in terms of a five-factor BART analysis and propose a determination for BART.

a. Control Technology Availability, Technical Feasibility, and Effectiveness

EPA proposes to find that SNCR and selective catalytic reduction (SCR) are available and technically feasible options to control NO

emissions with a control efficiency of approximately 50 percent for SNCR and approximately 89 percent for SCR.

SNCR involves the non-catalytic decomposition of NO

to molecular nitrogen and water. Typical NO

control efficiencies for SNCR range from 40 to 60 percent, depending on inlet NO

concentrations, fluctuating flue gas temperatures, residence time, amount and type of nitrogenous reducing agent, mixing effectiveness, acceptable levels of ammonia slip, and presence of interfering chemical substances in the gas stream. Because Sundt Unit 4 already operates with NO

combustion controls, we have used an SNCR control efficiency of 30 percent from a baseline that includes LNB with OFA. Considering typical combustion control technologies such as LNB and OFA can achieve control efficiencies of about 25 to 30 percent, the result is total control efficiency from an uncontrolled baseline of about 50 percent, which is in the mid-range of SNCR control efficiencies.

SCR is a post-combustion gas treatment technique that uses either ammonia or urea in the presence of a metal-based catalyst to selectively reduce NO

to molecular nitrogen, water, and oxygen. The catalyst lowers the temperature required for the chemical reaction between NO

and the reducing agent. Technical factors that impact the effectiveness of this technology include the catalyst reactor design, operating temperature, type of fuel fired, sulfur content of the fuel, design of the ammonia injection system, and the potential for catalyst poisoning. SCR has been installed on numerous coal-fired boilers of varying sizes, and is considered technically feasible. We note that SCRs are classified as a low dust SCR (LDSCR) or high dust SCR (HDSCR). As explained in the TSD, the SCR system considered in this analysis is the HDSCR.

Existing vendor literature and technical studies indicate that SCR systems are capable of achieving approximately 80 to 90 percent control efficiency, and that this emission rate can be achieved on a retrofit basis, particularly when combined with combustion control technology such as LNB.

Our contractor used a design emission rate of 0.050 lb/MMBtu (annual average), which in the case of Sundt Unit 4 corresponds to a control efficiency of 89 percent. While this is a value close to the upper range of SCR control efficiency, we consider the use of 0.050 lb/MMBtu appropriate for Sundt Unit 4. A review of Acid Rain Program data indicates that there are up to seven dry-bottom, wall-fired boilers operating with SCR on a retrofit basis that have achieved an annual average emission rate of 0.050 lb/MMBtu or lower in practice.

However, there are design differences between Sundt Unit 4 and these other units (i.e., boiler size, coal type and characteristics, and loading profile) that have the potential to affect this comparison. If we receive additional comments that sufficiently document source-specific considerations justifying the use of an emission rate higher than 0.050 lb/MMBtu, we may incorporate such considerations in our selection of BART.

See “Emissions Control: Cost-Effective Layered Technology for Ultra-Low NOx Control” (2007), “What's New in SCRs” (2006), and “Nitrogen Oxides Emission Control Options for Coal-Fired Electric Utility Boilers” (2005).

See spreadsheet “CAMD Wall-fired Coal EGUs.xlsx” in the docket.

b. BART Analysis for NO

Costs of Compliance:

In evaluating the costs of compliance for SNCR and SCR, we calculated the control costs ($) and emission reductions (tons/year of pollutant) for each control technology, and developed average cost-effectiveness ($/ton) values. Estimated NO

emission reductions are summarized in Table 7 and cost-effectiveness numbers are summarized in Table 8 for each option. A more detailed version of emission calculations are in our docket

and in our contractor's report. The heat duty and capacity factor used in the emission calculations below differ from the values used in the calculations originally prepared by our contractor, due to the unit's lower capacity when burning coal (130 MW) rather than natural gas (173 MW). The heat duty (MMBtu/hr) and capacity factor (0.49) reflect the coal-burning heat duty, rather than the natural gas-burning heat duty.

See spreadsheet “Sundt4 2001-12 Emission Calcs 2014-01-24.xlsx” in the docket.

As noted by TEP in its May 10, 2013 letter, although the calculated capacity factor is different, the annual emissions in tons per year removed do not change significantly, as the change in capacity factor is largely offset by the change in maximum unit gross rating.

Table 7—Sundt 4: NO

Control Option Emission Estimates

Control option

Control

efficiency

Emission

factor

lb/MMBtu

Heat duty

MMBtu/hr

Capacity factor

emission rate

lb/hr

tpy

emission

reduction

tpy

Baseline (LNB+OFA)

0.445

1,371

0.49

610

1,310

SNCR+LNB+OFA

0.312

1,371

0.49

427

917

393

SCR+LNB+OFA

0.050

1,371

0.49

147

1,162

Our consideration of the cost of compliance focuses primarily on the cost-effectiveness of each control option as measured in average cost per ton and incremental cost per ton of each control option as shown in Table 8. SCR is the most stringent option with the highest average cost-effectiveness of $5,176/ton, and incremental cost-effectiveness over SNCR of $6,174/ton. Detailed cost calculations can be found in our docket.

While we have relied primarily upon the cost calculations prepared by our contractor, we have incorporated certain elements of TEP's analysis

into our cost calculations. The most significant revisions to cost estimates include the following:

See spreadsheet “Sundt4 Control Costs 2014-01-26.xlsx” in the docket.

Letter dated May 10, 2013.

• We have changed the unit size from 173 MW to 130 MW to reflect the gross capacity of using coal. Although this has the net effect of decreasing certain costs, particularly several operation and maintenance (O&M) costs, the revised capital cost estimates increased for SCR (from $38 million to $45 million) and SNCR (from $2.8 million to $3.1 million).

• We have used a retrofit difficulty value of 1.5 (increased from 1.0) in cost estimates due to certain difficulties associated with retrofit installation of SCR. These difficulties are the result of site congestion and the configuration of the existing boiler structure and coal handling system as noted by TEP.

• We have included the cost of air preheater modifications that TEP stated are necessary in order to accommodate SCR due to site congestion and coal handling configuration.

Table 8—Sundt 4: NO

Control Option Cost-Effectiveness

Control option

Capital

cost

($)

Annualized capital cost

($)

Annual operating cost

($)

Total annual cost

($/yr)

Emission

reduction

(tpy)

Cost-effectiveness

($/ton)

Ave

Incremental

SNCR

$3,079,089

$290,644

$975,124

$1,265,768

393

$3,222

SCR

45,167,561

4,263,498

1,753,975

6,017,474

1,162

5,176

$6,174

Pollution Control Equipment in Use at the Source:

The presence of existing pollution control technology at Sundt Unit 4 is reflected in the consideration of available control technologies and in the development of baseline emission rates for use in cost calculations and visibility modeling. In the case of NO

current pollution controls are reflected in our selection of 2011 as the baseline period, which includes the use of LNB and OFA.

Energy and Non-Air Quality Environmental Impacts:

Regarding potential energy impacts of the BART control options, we note that SCR incurs a draft loss that will result in certain load loss, and that other emissions controls may also have modest energy impacts. The costs for direct energy impacts, i.e., power consumption from the control equipment and additional draft system fans from each control technology, are included in the cost analyses. Indirect energy impacts, such as the energy to produce raw materials, are not considered, which is consistent with the BART Guidelines. Ammonia adsorption (resulting from ammonia injection from SCR or SNCR) to fly ash is generally not desirable due to odor but does not impact the integrity of the use of fly ash in concrete. The ability to sell fly ash is unlikely to be affected by the installation of SNCR or SCR technologies. Finally, SNCR and SCR may involve potential safety hazards associated with the transportation and handling of anhydrous ammonia. However, since the handling of anhydrous ammonia will involve the development of a risk management plan (RMP), we consider the associated safety issues to be manageable as long as established safety procedures are followed. As a result, we do not consider these impacts sufficient to warrant the elimination of either of the available control technologies.

Remaining Useful Life of the Source:

We are considering the “remaining useful life” of Sundt Unit 4 as one element of the overall cost analysis as allowed by the BART Guidelines.

Since there is not state- or federally-enforceable shut-down date for this unit, we have used a 20-year amortization period per the EPA Cost Control Manual as the remaining useful life for the facility.

40 CFR Part 51, Appendix Y, section IV.D.4.k.

We note that the 20 year amortization period is primarily used in NO

control cost calculations, such as for SCR. In order to promote consistency in the analysis, we have used the 20 year period in the cost calculations for other control options, such as for SO

control, for which the Control Cost Manual includes examples that use an amortization period of 15 years.

Degree of Visibility Improvement:

The visibility improvement due to NO

controls is modest. SNCR was modeled at a 30 percent NO

emission reduction. As shown in Table 9, this yields a maximum visibility improvement of just over 0.2 dv at Saguaro. Galiuro improves about half as much, and other areas much less. The cumulative improvement across all impacted Class I areas is 0.5 dv. SCR was modeled at 89 percent NO

reduction to achieve 0.05 lb/MMBtu. SCR provides a maximum improvement of 0.8 dv, which occurs at Saguaro. Galiuro again improves about half as much, and the cumulative improvement across all Class I areas is 1.6 dv. This visibility improvement is substantially greater for SCR than for SNCR.

Table 9—Sundt 4: Visibility Impact and Improvement From NO

Controls

Class I area

Distance (km)

Visibility

impact

Base case

Visibility

improvement

SNCR

(ctrl04)

SCR

(ctrl08)

Chiricahua NM

144

0.43

0.03

0.12

Chiricahua WA

141

0.51

0.05

0.15

Galiuro WA

1.10

0.12

0.34

Gila WA

232

0.17

0.02

0.04

Mazatzal WA

203

0.19

0.02

0.04

Mount Baldy WA

232

0.15

0.01

0.03

Pine Mountain WA

247

0.15

0.02

0.03

Saguaro NP

3.40

0.23

0.78

Sierra Ancha WA

178

0.19

0.01

0.04

Superstition WA

137

0.32

0.01

0.05

Cumulative (sum)

6.6

0.5

1.6

Maximum

3.40

0.23

0.78

# CIAs >= 0.5 dv

Million $/dv (cumul. dv)

$2.4

$3.7

Million $/dv (max. dv)

$5.5

$7.7

c. Proposed BART Determination for NO

EPA proposes to find that BART for NO

is an emission limit of 0.36 lb/MMBtu on a 30-day BOD rolling basis that is achievable by SNCR with LNB and OFA. The primary factors supporting this proposed finding are the average cost-effectiveness and anticipated visibility benefits of controls. In particular, while SCR is anticipated to achieve the greatest degree of visibility improvement, it is also significantly more expensive than SNCR, with an average cost-effectiveness of $5176/ton. We do not consider this average cost to be warranted by the projected visibility benefit of SCR for this facility. Table 10 provides a summary of our five-factor BART analysis.

In proposing an emission limit of 0.36 lb/MMBtu, we have considered the annual average design value for SNCR of 0.31 lb/MMBtu as well as the need to account for emissions associated with startup and shutdown events. To account for this variability, we have examined the difference between the highest 30-day rolling NO

value and the highest annual average NO

value observed over the baseline period, which is approximately 17 percent.

We have applied this variability to the annual average design value to develop a 30-day BOD rolling emission limit, which we consider to provide sufficient margin for a limit that will apply at all times.

See spreadsheet “Sundt4 2001-12 Emission Calcs 2014-01-24.xlsx” in the docket.

We propose to require compliance with this requirement within three years of the effective date of the final rule. A 2006 Institute of Clean Air Companies (ICAC) study indicated that the installation time for a typical SNCR retrofit, from bid to startup, is 10 to 13 months.

However, because we are also requiring the installation of additional SO

controls, we consider a three year period for compliance with both BART determinations to be appropriate. We are seeking comment on whether this compliance date is reasonable and consistent with the requirement of the Clean Air Act that BART be installed “as expeditiously as practicable but in no event later than five years after [promulgation of the applicable FIP].”

If we receive information during the comment period that establishes that a different compliance time frame is appropriate, we may finalize a different compliance date. Finally, we are proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements to ensure that the emission limit and compliance deadline are enforceable. As part of the proposed monitoring requirements, we are including a requirement to monitor rates of ammonia injection in order to ensure proper operation of the SNCR in a manner that minimizes ammonia emissions.

See “Typical Installation Timelines for NO

Emissions Control Technologies on Industrial Sources”, Institute of Clean Air Companies, December 4, 2006.

Clean Air Act section 169A(g)(4), 42 U.S.C. 7491(g)(4).

Table 10—Sundt 4: Summary of BART Analysis for NO

Sundt unit 4 (130 MW)

LNB+OFA

(baseline)

SNCR+LNB

SCR+LNB

Emissions

Emission Factor (lb/MMBtu)

0.445

0.312

0.050

Emission Rate (tpy)

1310

917

147

Emission Reduction (tpy)

393

1,162

Control Effectiveness (%)

30%

89%

Costs of Compliance

Capital Cost ($)

$3,079,089

$45,167,561

Annualized Capital Cost ($)

$290,644

$4,263,498

Annual O&M ($)

$975,124

$1,753,975

Total Annual Cost ($)

$1,265,768

$6,017,474

Ave Cost-Effectiveness ($/ton)

$3,222

$5,176

Incremental Cost-Effectiveness ($/ton)

$6,174

Pollution Control Equipment in Use

Low-NO

Burners and Over Fire Air

Energy and Non-Air Quality Environmental Impacts

Energy impacts have been reflected in annual O&M costs in the costs of compliance.

SCR and SNCR may create potential safety and environmental hazards from the transportation and handling of anhydrous ammonia. We consider these impacts manageable with the development of an RMP and additional safety procedures, and do not consider them sufficient enough to warrant eliminating either of these available control technologies.

Remaining Useful Life

Control technology amortization period

20 years

Visibility Improvement

Single largest Class I area improvement (dv)

0.23

0.78

Single Class I area cost-effectiveness (million $/dv)

$5.5

$7.7

Class I areas with ≥ 0.50 dv improvement

Cumulative visibility improvement (dv)

0.5

1.6

Cumulative cost-effectiveness (million $/dv)

$2.4

$3.7

4. Proposed BART Analysis and Determination for SO

For our SO

BART analysis, we identified all available control technologies, eliminated options that are not technically feasible, and evaluated the control effectiveness of the remaining control options. We then evaluated each control in terms of a five-factor BART analysis and proposed a determination for BART.

a. Control Technology Availability, Technical Feasibility, and Effectiveness

EPA identified three available and technically feasible technologies to control SO

emissions from Sundt Unit 4. These technologies are lime or limestone-based wet flue gas desulfurization (wet FGD), lime spray dry absorber (SDA or dry FGD), and dry sorbent injection (DSI). While each of these control options has certain design concerns and constraints associated with their implementation, all three options are considered technically feasible.

Lime or limestone-based wet FGD:

Wet scrubbing systems mix an alkaline reagent, such as hydrated lime or limestone, with water to generate scrubbing slurry that is used to remove SO

from the flue gas. The alkaline slurry is sprayed countercurrent to the flue gas, such as in a spray tower, or the flue gas may be bubbled through the alkaline slurry as in a jet bubbling reactor. As the alkaline slurry contacts the exhaust stream, it reacts with the SO

in the flue gas. Design variations may include changes to increase the alkalinity of the scrubber slurry, increase slurry/SO

contact, and minimize scaling and equipment problems. Insoluble calcium sulfite (CaSO

) and calcium sulfate (CaSO

) salts are formed in the chemical reaction that occurs in the scrubber, and exit as part of the scrubber slurry. The salts are eventually removed and handled as a solid waste byproduct. The waste byproduct is mainly CaSO

which is difficult to dewater. Solid waste byproducts from wet lime scrubbing are typically managed in dewatering ponds and landfills.

Design concerns associated with wet FGD involve the substantial water usage requirements needed to generate the alkaline reagent slurry as well as the substantial amount of wastewater and solid waste discharge associated with the spent byproduct. A wet FGD control system must be located after the fabric filter baghouse because the moist plume resulting from the wet scrubber system would create baghouse plugging issues if the control is placed ahead of the baghouse. In addition, a substantial footprint is required for the management of these waste products as well as for the absorber tower and associated process equipment such as the slurry preparation, mixing, associated tanks, and dewatering activities. While these design concerns do present some challenges, they do not warrant elimination of this option as technically infeasible.

TEP's review does not eliminate consideration of wet FGD, but does describe several design challenges that TEP notes should be reflected in the five factor analysis. We have incorporated certain elements of TEP's review in our analysis, as discussed in Step 4.

Our contractor has estimated that newly constructed wet FGD systems could achieve design emission rates (annual average basis) of 0.06 lb/MMBtu. Relative to baseline SO

emission rates, this corresponds to a control efficiency of 92 percent. We recognize that FGD systems are designed to achieve more stringent emission rates, and have demonstrated an ability to achieve control efficiencies up to 98 percent. Our contractor's report notes that the lower control efficiency cited here is regarded as a conservative estimate. While this is not the most stringent level of control that the technology is capable of achieving, we consider 92 percent control efficiency to be consistent with the median values reported for wet FGD systems.

Lime SDA or dry FGD:

A spray dryer absorber uses a stream of either dry lime or hydrated lime (semi-dry) in a reaction tower where it reacts with SO

in the flue gas to form calcium sulfite solids. Unlike wet FGD systems that produce a slurry by-product that is collected

separately from the fly ash, dry FGD systems are designed to produce a dry byproduct that must be removed with the fly ash in the particulate control equipment. As a result, dry FGD systems must be located upstream of the particulate control device to remove the reaction products and excess reactant material. In instances where hydrated lime is used as a reagent, the reaction towers must be designed to provide adequate contact and residence time between the exhaust gas and the slurry to produce a relatively dry byproduct. Typical process equipment associated with a spray dryer typically includes an alkaline storage tank, mixing and feed tanks, an atomizer, spray chamber, particulate control device and a recycle system. The recycle system collects solid reaction products and recycles them back to the spray dryer feed system to reduce alkaline sorbent use.

A design concern associated with a dry FGD system is that it must be installed prior to the fabric filter baghouse in order for the reagent to be captured and recycled. As noted in our contractor's report, the location of the existing fabric filter baghouse does not present enough space to install a new absorber between the boiler and the existing baghouse. As a result, a dry FGD at Sundt Unit 4 is assumed to include a new baghouse, which is reflected in the costs of compliance for the five-factor analysis. We consider this control option to be technically feasible.

Our contractor has estimated that newly constructed dry FGD systems could achieve design emission rate (annual average basis) of 0.08 lb/MMBtu. Relative to baseline SO

emission rates, this corresponds to a control efficiency of 89 percent. As noted for wet FGD systems, this is a conservative estimate of what dry FGD systems can achieve, and is consistent with the median values reported for dry FGD systems.

Dry Sorbent Injection:

DSI involves the injection of powdered absorbent directly into the flue gas exhaust stream. These are simple systems that generally require a sorbent storage tank, feeding mechanism, transfer line and blower, and an injection device. The dry sorbent is typically injected countercurrent to the gas flow. An expansion chamber is often located downstream of the injection point to increase residence time and efficiency. Particulates generated in the reaction are controlled in the system's particulate control device. DSI requires less capital equipment, less physical space, and less modification to existing ductwork compared to a dry FGD system. However, reagent costs are much higher and, depending upon the absorbent and amount of sorbent injected, control efficiency is lower when compared to a dry FGD system. Soda ash and Trona (sodium sesquicarbonate) are potential options for reagent use. An important design consideration of DSI is the ability of the downstream particulate control device to accommodate the additional particulate loading resulting from the addition of the DSI reagent into the boiler flue gas. More effective particulate control devices allow for higher rates of sorbent injection, which in turn allow for more effective SO

control.

In a review of SO

control options for BART eligible units, the Northeast States for Coordinated Air Use Management (NESCAUM) estimated control effectiveness for DSI in a range of 40-60 percent.

More recently, as part of work done as part of the Integrated Planning Model (IPM), EPA has estimated control effectiveness as high as 80 percent,

depending upon factors such as the type of sorbent, the quantity of sorbent used, and the type of particulate control device employed. Generally, the use of more effective particulate control devices allow for higher rates of sorbent injection, and therefore greater DSI effectiveness. Since Sundt Unit 4 operates with a fabric filter, we consider a control effectiveness value in the upper range appropriate, and have used 70 percent control effectiveness in our calculations. This value is above the range indicated in the NESCAUM study, but does not require the high sorbent injection rates required to achieve the upper range of control indicated in IPM documentation. A summary of the control technologies and their associated control effectiveness is presented in Table 11.

“Assessment of Control Technology Options for BART-Eligible Sources”, Northeast States for Coordinated Air Use Management In Partnership with The Mid-Atlantic/Northeast Visibility Union, March 2005.

IPM Model—Revisions to Cost and Performance for APC Technologies, Dry Sorbent Injection Cost Development Methodology, August 2010.

Table 11—Sundt 4: SO

Control Options

Control option

Control effectiveness

Dry Sorbent Injection

Dry FGD or Lime SDA

Wet FGD (lime- or limestone-based)

b. BART Analysis for SO

Costs of Compliance:

Our consideration of the costs of compliance focuses primarily on the cost-effectiveness of each control option, as measured in cost per ton and incremental cost per ton. The emissions estimates and cost-effectiveness for the three control options are shown in Table 12 and Table 13, respectively. Both wet and dry FGD have average cost-effectiveness values over $5,000/ton, much greater than DSI, which is a control option that we consider very cost-effective at $1,857/ton. Moreover, both wet and dry FGD have very high incremental cost-effectiveness values, indicating that while they are more effective than less stringent control options, this additional degree of effectiveness comes at a substantial cost.

In evaluating the costs of compliance for the control options, we have calculated the control costs ($) and emission reductions (tons/year of pollutant) for each control technology, developed average cost-effectiveness ($/ton) values, and arrived at the emission reductions for each option as summarized Table 12. A more detailed version of emission calculations is in our docket,

and in our contractor's report. As noted previously in our NO

BART analysis, the heat duty and capacity factor used in these calculations differ from the values used in the calculations originally prepared by our contractor because the maximum gross capacity of Sundt Unit 4 while burning coal is about 130 MW, compared to its natural-gas nameplate capacity of 173 MW. The heat duty (MMBtu/hr) and capacity factor used in Table 12 reflect the coal-burning nameplate capacity.

Detailed cost calculations presented in Table 13 are in the docket.

See spreadsheet “Sundt4 2001-12 Emission Calcs 2014-01-24.xlsx” in the docket.

As noted by TEP and Burns and McDonnell, although the calculated capacity factor is different, the annual emissions in tons per year removed do not change significantly, as the change in capacity factor is largely offset by the change in maximum unit gross rating.

See spreadsheet “Sundt4 Control Costs 2014-01-26.xlsx” in the docket.

Table 12—Sundt 4: SO

Control Option Emission Estimates

Control option

Control

efficiency

(%)

Emission factor

(lb/MMBtu)

Heat

duty

(MMBtu/hr)

Capacity factor

emission rate

(lb/hr)

(tpy)

emission

reduction

(tpy)

Baseline (no control)

0.729

1,371

0.49

1,000

2,145

DSI

0.219

1,371

0.49

300

644

1,502

DFGD

0.080

1,371

0.49

110

236

1,909

WFGD

0.060

1,371

0.49

177

1,969

Table 13—Sundt 4: SO

Control Option Cost-Effectiveness

Control option

Capital

cost

($)

Annualized capital cost

($)

Annual operating cost

($)

Total annual cost

($/yr)

Emission

reduction

(tpy)

Cost-effectiveness

($/ton)

Ave

Incremental

DSI

$3,250,000

$306,777

$2,482,107

$2,788,884

1,502

$1,857

DFGD

72,470,559

6,840,708

2,880,841

9,721,549

1,909

5,091

$17,007

WFGD

80,629,663

7,610,870

3,227,467

10,838,337

1,969

5,505

18,795

Pollution Control Equipment in use at Source:

In the case of SO

Sundt Unit 4 does not operate with any existing control technology. This is reflected in our selection of calendar year 2011 as the baseline period, which represents uncontrolled coal-fired emissions.

Energy and Non-Air Quality Environmental Impacts:

For wet FGD, energy impacts include certain auxiliary power requirements that are necessary to operate the wet FGD system and to potentially compensate for pressure head loss through the scrubber. These energy impacts are reflected as auxiliary power costs in the cost of compliance estimates. Non-air quality environmental impacts include water usage requirements and the storage and disposal of wet ash. Wet FGD requires very large quantities of water to ensure proper control effectiveness. Securing such quantities of water is a significant challenge in more arid regions of the country such as Arizona, and would preclude the use of that water for potentially more beneficial uses. The on-site storage and disposal of wet ash in large retention ponds triggers significant additional regulatory requirements, as it represents a substantial water pollution threat.

For dry FGD, the energy and non-air environmental impacts are similar to those for wet FGD. Operation of a dry FGD system still requires securing significant supplies of water, although to a lesser degree than wet FGD systems. In addition, dry FGD systems will result in generation of larger quantities of boiler ash, and has the potential to affect negatively the properties and quality of boiler ash. In some instances, boiler ash that is suitable to sell for beneficial purposes may no longer be marketable following installation of a dry FGD system. Energy impacts also include auxiliary power requirements for operation of the dry FGD system, and for overcoming pressure head loss through the scrubber. While we note certain potential impacts resulting from the water resource requirements associated with wet FGD as well as the additional solid waste generation associated with wet and dry FGD, we do not consider these impacts sufficient enough to warrant eliminating these control technologies.

DSI could potentially have an adverse effect on the quality of the boiler fly ash, which would make it unmarketable for beneficial uses. Use of DSI also results in an ash byproduct which would require landfill disposal, thereby increasing solid waste generation rates at the plant. Energy impacts are limited to auxiliary power requirements for operation of the DSI system. We do not consider these impacts sufficient enough to warrant eliminating this control technology.

Remaining Useful Life of the Source:

We are considering the remaining useful life of Sundt Unit 4 as one element of the overall cost analysis as allowed by the BART Guidelines. Since we are not aware of any federally- or State-enforceable shut down date for Sundt Unit 4, we have used a 20-year amortization period described in the EPA Cost Control Manual as the remaining useful life for the control options considered for Unit 4. We note that the remaining useful life of the source is reflected in the evaluation of cost of compliance through the use of a 20-year amortization period in control cost calculations.

Degree of Visibility Improvement:

The visibility improvement due to SO

controls is modest. As shown in Table 14, control via DSI, with a 70 percent SO

emissions reduction, gives a maximum visibility improvement of 0.2 dv, which occurs at Saguaro. Three other areas improve about half as much, and the cumulative improvement is 0.8 dv. Emissions controls via dry and wet FGD were modeled at 89 percent and 92 percent SO

emissions reduction, respectively. Both dry and wet FGD would cause a visibility disbenefit at Saguaro as indicated by the negative improvements in Table 14. The disbenefit is mainly due to the decreased stack exit temperature and exit velocity associated with these technologies, and more so for wet FGD than for dry FGD. These stack decreases result in less plume rise and increased impacts nearby. At areas farther away, the disbenefit is outweighed by the benefit of SO

reductions from FGD. This issue is discussed further in the TSD. With FGD, the maximum benefit occurs not at Saguaro, but at Galiuro, with 0.2 dv for dry FGD and 0.1 dv for wet FGD. The corresponding cumulative improvements are 0.6 dv and 0.4 dv for dry and wet FGD, respectively, including the areas of disbenefit. All these improvements are substantially lower than those from DSI, and the visibility cost-effectiveness of each FGD is more than quadruple that of DSI. EPA finds that the improvement from DSI is substantial enough to support its selection as BART, and that it is clearly a better choice than dry FGD and wet FGD.

Table 14—Sundt 4: Visibility Impact and Improvement From SO

Controls

Class I Area

Distance (km)

Visibility impact

Base

case

Visibility improvement

DSI 70%

(ctrl14)

Dry FGD

(ctrl02)

Wet FGD

(ctrl03)

Chiricahua NM

144

0.43

0.05

0.07

0.06

Chiricahua Wild

141

0.51

0.10

0.11

Galiuro Wild

1.10

0.10

0.16

0.09

Gila Wild

232

0.17

0.04

0.05

Mazatzal Wild

203

0.19

0.07

0.08

0.09

Mount Baldy Wild

232

0.15

0.05

0.06

Pine Mountain Wild

247

0.15

0.05

0.06

Saguaro NP

3.40

0.20

−0.16

−0.27

Sierra Ancha Wild

178

0.19

0.06

0.08

Superstition Wild

137

0.32

0.09

0.10

Cumulative (sum)

6.6

0.8

0.6

0.4

Maximum

3.40

0.20

0.16

0.11

# CIAs >= 0.5 dv

Million $/dv (cumul. dv)

$3.5

$16.4

$25.1

Million $/dv (max. dv)

$14

$60

$97

c. BART Determination for SO

EPA proposes an emission limit of 0.23 lb/MMBtu on a 30-day (BOD) rolling basis as BART to control SO

from Sundt Unit 4. This emission limit, equivalent to using DSI, is considered very cost-effective at $1,857/ton. In evaluating the appropriate emission limit for DSI, we have considered the annual average design value for DSI of 0.21 lb/MMBtu as well as the need to account for emissions associated with startup and shutdown events. To determine how to account for this variability, we have examined the difference between the highest 30-day rolling SO

value and the highest annual average SO

value observed over the baseline period, which is approximately 9 percent.

We have applied this variability to the annual average design value to develop a 30-day BOD rolling emission limit, which we consider a sufficient margin for a limit that will apply at all times. Please refer to Table 15 that provides a summary of our five-factor BART analysis.

See spreadsheet “Sundt4 2001-12 Emission Calcs 2014-01-24.xlsx” in the docket.

We propose to require compliance with this requirement within three years of the effective date of the final rule. However, we are seeking comment on whether this compliance date is reasonable and consistent with the requirement of the Clean Air Act that BART be installed “as expeditiously as practicable but in no event later than five years after [promulgation of the applicable FIP].”

If we receive information during the comment period that establishes that a different compliance time frame is appropriate, we may finalize a different compliance date. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this emission limit.

Clean Air Act section 169A(g)(4), 42 U.S.C. 7491(g)(4).

Table 15—Sundt 4: Summary of BART Analysis for SO

Sundt Unit 4 (130 MW)

Baseline

DSI

Dry FGD

Wet FGD

Emission Factor (lb/MMBtu)

0.729

0.219

0.08

0.06

Emission Rate (tpy)

2145

644

236

177

Emission Reduction (tpy)

1,502

1,909

1,969

Control Effectiveness

70%

89%

92%

Cost of Compliance

Capital Cost ($)

$3,250,000

$72,470,559

$80,629,663

Annualized Capital Cost ($)

$306,777

$6,840,708

$7,610,870

Annual O&M ($)

$2,482,107

$2,880,841

$3,227,467

Total Annual Cost ($)

$2,788,884

$9,721,549

$10,838,337

Ave CE ($/ton)

$1,857

$5,091

$5,505

Incremental CE ($/ton)

$23,081

$18,795

Pollution Control Equipment in Use at Source

There is no existing control technology for SO

Energy and Non-Air Quality Environmental Impacts

Energy impacts are reflected in annual O&M costs in the costs of compliance.

Wet ash from wet and dry FGD represents a substantial water pollution threat.

Water resources for wet and dry FGD may preclude more beneficial uses of water.

Remaining Useful Life

Control technology amortization period

20 years

Visibility Improvement

Single largest Class I area improvement (dv)

0.20

0.16

0.11

Single Class I area cost-effectiveness (million $/dv)

$14.3

$60.4

$96.8

Class I areas with ≥ 0.50 dv improvement

Cumulative visibility improvement (dv)

0.8

0.6

0.4

Cumulative cost-effectiveness (million $/dv)

$3.5

$16.4

$25.1

3. Proposed BART Analysis and Determination for PM

a. Control Technology Availability, Technical Feasibility, and Effectiveness

Sundt Unit 4 currently operates with a fabric filter baghouse for particulate control, which is considered the most stringent control device for particulate matter. These devices operate on the same principle as a vacuum cleaner. Air carrying dust particles is forced through a cloth bag that is designed and manufactured to trap particles greater than a certain specified diameter. As the air passes through the fabric, the dust accumulates on the cloth and is removed from the air stream. The accumulated dust is periodically removed from the cloth by shaking or by reversing the air flow. The layer of dust, known as dust cake, trapped on the surface of the fabric has the potential to result in high efficiency rates for particles ranging in size from submicron to several hundred microns in diameter.

b. BART Analysis for PM

The BART Guidelines provide that, where a source has controls already in place that are the most stringent controls available, it is not necessary to complete comprehensively a full five-factor BART analysis, as long the most stringent controls available are made federally enforceable. Therefore, instead of completing the remaining steps of a five-factor BART analysis, we have evaluated the appropriate level of emissions to ensure that the fabric filter achieves an appropriate degree of control.

c. Proposed BART Determination for PM

EPA is proposing a filterable PM

BART emission limit of 0.03 lb/MMBtu based on the use of the existing fabric filter baghouse currently in operation, which is the most stringent control for particulate matter. We note that Mercury and Air Toxics (MATS) Rule establishes an emission standard of 0.03 lb/MMBtu filterable PM (as a surrogate for toxic non-mercury metals) as representing Maximum Achievable Control Technology (MACT) for coal-fired EGUs.

This standard derives from the average emission limitation achieved by the best performing 12 percent of existing coal-fired EGUs, as based upon test data used in developing the MATS Rule.

The BART Guidelines provide that, “unless there are new technologies subsequent to the MACT standards which would lead to cost-effective increases in the level of control, you may rely on the MACT standards for purposes of BART.”

Therefore, we propose to find that 0.03 lb/MMBtu filterable PM

is an appropriate limit for BART at Sundt Unit 4.

77 FR 9304, 9450, 9458 (February 16, 2012) (codified at 40 CFR 60.42Da(a), 60.50Da(b)(1)).

See Memorandum from Jeffrey Cole (RTI International) to Bill Maxwell (EPA) regarding “National Emission Standards for Hazardous Air Pollutants (NESHAP) Maximum Achievable Control Technology (MACT) Floor Analysis for Coal- and Oil-fired Electric Utility Steam Generating Units for Final Rule” (December 16, 2011).

40 CFR Part 51, Appendix Y, Section IV.C.

4. Better Than BART Alternative

We are proposing a switch to natural gas on Sundt Unit 4 as a better-than-BART alternative to the emissions controls previously proposed in this section for a coal-fired unit. Unit 4 was originally constructed as a natural gas-fired boiler, and has used natural gas as a primary fuel for significant periods of time since 2009. While a change in fuel supply to natural gas instead of coal is an inherently less polluting option, the BART Guidelines do not require the consideration of fuel supply changes as a control option.

As a result, the option of burning only natural gas is not considered in our BART analysis. However, TEP has submitted to EPA an alternative to BART based on the elimination of coal as a fuel source for Sundt Unit 4 by December 31, 2017. As part of this submittal, TEP compared the potential emission reductions and visibility benefit between a natural gas fuel change and certain combinations of NO

and SO

controls.

40 CFR Part 51, Appendix Y, Section IV.D.1.5, “STEP 1: How do I identify all available retrofit emission control techniques?”

Letter dated November 1, 2013.

EPA has evaluated this alternative proposal pursuant to the “better-than-BART” provisions of the RHR. In particular, the RHR allows for implementation of “an emissions trading program or other alternative measure” in lieu of BART if the alternative measure achieves greater reasonable progress than would be achieved through the installation and operation of BART.

The rule further states that “[i]f the distribution of emissions is not substantially different than under BART, and the alternative measure results in greater emissions reductions, than the alternative measures may be deemed to achieve greater reasonable progress”.

Because the emissions reductions under EPA's BART proposal for Sundt Unit 4 and the reductions from TEP's proposed alternative would occur at the same facility, the distribution of emissions under BART and the alternative are not substantially different. Therefore, if the alternative emission control strategy results in greater emissions reductions than our BART proposal, EPA may deem the alternative emission control strategy to achieve greater reasonable progress. A comparison of annual emission estimates between the BART determination and alternative to BART is summarized in Table 16. BART determination annual emissions are based upon the annual average emission factors and annual capacity factor used in our BART analysis, consistent with coal usage. For the alternative to BART, annual emissions are based on a combination of historical natural gas usage data as indicated in TEP's submittal, as well as standard emission factors for natural gas combustion. A more detailed discussion of emission estimates from these two scenarios is included in our TSD.

40 CFR 51.308(e)(2).

40 CFR 51.308(e)(3).

Table 16—Sundt 4: Comparison of BART Determination and Alternative to BART

Parameters

Units

BART determination

Natural gas fuel switch

Difference

Heat Duty

MMBtu/hr

1,371

1,828

Capacity Factor

0.49

0.37

Ctrl Tech

SNCR+LNB+OFA

LNB+OFA

lb/MMBtu

0.31

0.22

tpy

917

652

265

Particulate Matter

Ctrl Tech

Fabric Filter

None

lb/MMBtu

0.03

0.01

tpy

Ctrl Tech

Dry Sorbent Injection

None

lb/MMBtu

0.22

0.00064

tpy

644

1.9

642

Annual average emission factors.

As seen in Table 16, a change to natural gas usage achieves greater emission reductions than each of the individual BART determinations for NO

, SO

, and particulate matter, as well as in the aggregate. Although visibility modeling is not required to support a better-than-BART determination in this instance, EPA conducted modeling to verify the visibility benefits of the proposed alternative, as compared with EPA's BART determination. This modeling is described in the TSD and the results are summarized in Table 17.

Table 17—Sundt 4: Visibility Impact and Improvement From Combined SO

and NO

BART, and From Better-Than-BART Alternative

Class I Area

Distance (km)

Visibility

impact

Base case

Visibility improvement

SNCR DSI 70%

(ctrl15)

Natural gas

(ctrl13)

Chiricahua NM

144

0.43

0.09

0.19

Chiricahua WA

141

0.51

0.16

0.25

Galiuro WA

1.10

0.24

0.47

Gila WA

232

0.17

0.06

0.10

Mazatzal WA

203

0.19

0.08

0.12

Mount Baldy WA

232

0.15

0.06

0.09

Pine Mountain WA

247

0.15

0.06

0.09

Saguaro NP

3.40

0.49

1.06

Sierra Ancha WA

178

0.19

0.08

0.12

Superstition WA

137

0.32

0.11

0.19

Cumulative (sum)

6.6

1.4

2.7

Maximum

3.40

0.49

1.06

# CIAs >= 0.5 dv

Million $/dv (cumul. dv)

$2.8

Million $/dv (max. dv)

$8.3

Since Sundt is only 17 km from the eastern unit of Saguaro, its emitted NO

may not be fully converted to NO

by the time it reaches there, as is assumed in the CALPUFF model. It thus may not be fully available to form visibility-degrading particulate nitrate. EPA explored this issue in CALPUFF sensitivity simulations described in the TSD. For EPA's proposed BART of SNCR plus DSI, the visibility improvement remains above 0.3 dv even when unrealistically low 10 percent NO-to-NO

conversion is assumed (i.e., no additional conversion of NO to NO

once the plume leaves the stack). The improvement from switching to natural gas remains above 0.7 dv at Saguaro. These results show that the FIP's proposed BART determination remains reasonable despite any concern over the NO conversion rate; the visibility improvement from BART remains substantial. The finding that natural gas provides better visibility improvement than the proposed BART determination also remains sound regardless of the NO conversion assumed.

Based on this information, we consider a natural gas fuel switch to result in greater emission reductions and achieve greater reasonable progress than the proposed BART determinations. Under this scenario, we are proposing a NO

emission limit of 0.25 lb/MMBtu based on a 30-day BOD rolling average. As discussed previously in the NO

BART determination, this represents about a 17 percent increase from the annual average emission rate of 0.22 lb/MMBtu, which we consider to provide sufficient margin for a limit that will apply at all times, including periods of startup and shutdown. In addition, we are proposing particulate matter and SO

emission limits consistent with natural gas use, as well as monitoring, reporting, and recordkeeping requirements.

B. Chemical Lime Nelson Plant Kilns 1 and 2

Summary:

EPA is proposing to find that Chemical Lime Nelson is subject to BART. EPA is proposing BART emission limits for NO

and PM

for Kilns 1 and 2 at the Nelson Plant as listed in Table 18 and described in this section.

Table 18—Nelson Lime Plant: Summary of Proposed BART Determinations

Source

Pollutant

Emission Limit (lb/ton feed)

Control technology* (for

reference only)

Kiln 1

3.80

Selective Non-Catalytic Reduction (SNCR).

9.32

Lower sulfur fuel.

0.12

Fabric filter baghouse (existing).

Kiln 2

2.61

Selective Non-Catalytic Reduction (SNCR).

9.73

Lower sulfur fuel.

0.12

Fabric filter baghouse (existing).

* The facility is not required to install the listed technology to meet the BART limit.

Affected Class I Areas:

Nine Class I areas are within 300 km of the Nelson Lime Plant. Their nearest borders range from 24 km to 289 km away, with the Grand Canyon the closest and other areas more than 100 km away. The highest baseline visibility impact from the Nelson Plant is 1.79 dv at Grand Canyon NP followed by 0.31 at Sycamore Canyon WA and 0.28 at Zion NP. The cumulative sum of visibility impacts over all the Class I areas is 3.34 dv.

Facility Overview:

The Nelson Plant processes limestone and manufactures lime near Peach Springs in Yavapai County, Arizona. The limestone processing plant consists of a quarry mining operation, a limestone crushing and screening operation, a limestone kiln feed system, a solid fuel handling system, two rotary lime kilns, front and back lime handling systems, a lime hydrator, diesel electric generators, fuel storage tanks, and other support operations and equipment. The lime manufacturing equipment consists of two lime rotary kilns (Kiln 1 and Kiln 2) and auxiliary equipment necessary for receiving crushed limestone, processing it through the lime kilns, and processing the lime kiln product. The lime kilns are used to convert crushed limestone (CaCO

) into quicklime (CaO).

We primarily relied on four sources of information for our proposed BART analyses and determinations. An initial BART analysis performed by our contractor

is available in the docket in the form of a final contractor's report and associated modeling spreadsheets. We also incorporated elements of a five-factor BART analysis

provided by Lhoist North America (LNA) of Arizona, owner of the Nelson Plant, that includes control cost estimates and visibility modeling. Another key document in our analysis is the Nelson Lime Plant's Title V Operating Permit.

Technical Analysis for Arizona and Hawaii Regional Haze FIPs: Task 7: Five-Factor BART Analysis for Chemical Lime Company Nelson, TEP Sundt (Irvington), and Catalyst Paper (Snowflake) Plants, Contract No. EP-D-07-102, Work Assignment 5-12; Prepared for EPA Region 9 by University of North Carolina at Chapel Hill, ICF International, and Andover Technology Partners; October 9, 2012.

BART Five Factor Analysis, Lhoist North America Nelson Lime Plant;

Prepared by Trinity Consultants in Conjunction with Lhoist North America of Arizona, Inc.; Project 131701.0061; August 2013. (Public version dated September 27, 2013).

Title V Operating Permit and Technical Support Document for the Nelson Lime Plant, Permit # 42782, Issued August 8, 2011 by the Arizona Department of Environmental Quality.

Baseline Emissions Calculations:

LNA's approach to establishing baseline emissions was to first establish baseline emission factors in lb/ton lime based on CEMS testing performed from March to June 2013. Annual average baseline emissions were calculated by multiplying these lb/ton emission factors by the highest annual lime production rate observed over a period from 2001 to 2012. Maximum daily emissions were calculated by multiplying lb/ton emission factors by the maximum daily lime production rate observed during the March to June 2013 testing period. As explained in further detail in our TSD, we consider LNA's general approach appropriate, but also note that it represents a conservatively high estimate of baseline emissions, and potentially overstates the anticipated emission reductions and visibility benefit from the evaluated control options. Nonetheless, given the lack of measured annual emissions data, we concur with LNA's use of a conservatively high baseline emissions estimate and we have incorporated this estimate into our analysis. The baseline daily and annual emission rates and associated production levels are shown in Table 19.

Table 19—Nelson Lime Plant: Summary of Maximum Daily and Annual Baseline Emissions for NO

and SO

Kiln

Lime production

Max daily

(tpd)

Max annual

(tpy)

Year

Emission factor

(lb/ton lime)

Maximum emissions

(lb/day)

(tpy)

Emission factor

(lb/ton lime)

Maximum emissions

(lb/day)

(tpy)

Kiln 1

866

258,508

2010

7.59

6,573

981

12.15

10,522

1,570

Kiln 2

1,246

378,296

2012

5.21

6,492

985

12.69

15,812

2,400

Maximum emission factors observed during March, May and June 2013 CEMS testing.

Maximum daily rates occurring during the March 2013 CEMS testing.

2010.

2012.

1. Proposed Subject to BART

As part of our July 30, 2013 final rulemaking on the Arizona RH SIP, we approved ADEQ's finding that Chemical Lime Nelson Plant (Nelson Lime Plant) Kilns 1 and 2 were BART-eligible, but disapproved ADEQ's determination that the Nelson Lime Plant was not subject to BART.

In light of this disapproval, we have conducted our own evaluation of whether Nelson Lime Plant is subject to BART, relying primarily on emissions

data and modeling results provided by the facility's owner, LNA.

78 FR 46175 (codified at 40 CFR 52.145(g)(1)(i)).

BART Five Factor Analysis, Lhoist North America Nelson Lime Plant; Prepared by Trinity Consultants in Conjunction with Lhoist North America of Arizona, Inc.; Project 131701.0061; August 13, 2013 (Public version dated September 27, 2013).

As explained in the TSD, the baseline emissions estimates and the corresponding modeling results provided by LNA are conservative (i.e., tending to overestimate rather than underestimate the impacts, in this case). Nonetheless, we consider these results to be appropriate for purposes of a subject-to-BART determination, as well as for the five-factor BART analysis. LNA's modeling results indicate that the 98th percentile impact for each of the 3 years modeled is well over 0.5 dv at Grand Canyon National Park.

Therefore, we propose to determine that Nelson Lime Plant (Kilns 1 and 2) is subject to BART.

Id.,

Table 4-7. We note that the visibility modeling performed by LNA used only the annual average Class I area background concentrations, rather than the best 20 percent days background concentrations. The use of annual average generally results in lower visibility impacts than the best 20 percent days. Therefore, had LNA used the best 20 percent days, the baseline impacts would likely have been even greater.

2. Proposed BART for NO

For our NO

a. Control Technology Availability, Technical Feasibility and Effectiveness

EPA proposes to find that SNCR is the only technically feasible control option to control NO

emissions with a control efficiency of 50 percent. In order to determine a reasonable performance standard for controlling NO

emissions, we considered four available retrofit control technologies for NO

on Kilns 1 and 2. These control technologies are a LNB, mixing air technology (MAT), SCR, and SNCR. After evaluating each of these technologies to eliminate technically infeasible options, we determined that SNCR is the only remaining technically feasible control option.

Low-NO

Burners:

LNB are designed to reduce flame turbulence, delay fuel/air mixing, and establish fuel-rich zones for initial combustion. LNA indicated that it experimented with the installation of bluff body LNB on the Nelson Lime Plant kilns in 2001.

These LNB wore out in about six months, negatively affected production, caused brick damage, and resulted in unscheduled shutdowns of the kilns. We recognize that the staged combustion principle of LNB can present operational difficulties and potential product quality issues for lime production that are not exhibited in the cement industry. At this time we consider LNB to be technically infeasible for the Nelson Plant kilns, since we do not have any information to suggest otherwise at this time. The technical feasibility of LNB will be re-evaluated for lime kilns in subsequent reasonable progress planning periods.

Described on page 5-2, “

BART Five Factor Analysis, Lhoist North America Nelson Lime Plant”

(Public version dated September 27, 2013).

Mixing Air Technology:

MAT is the practice of injecting a high pressure air stream into the middle of a kiln to help mix the air flowing through the kiln. While the theory behind MAT suggests that the technology is effective at reducing NO

emissions, it is not clear whether this control technology is effective on lime kilns. We propose to eliminate MAT as not technically feasible for retrofit on Kiln 1 and Kiln 2.

Selective Catalytic Reduction:

This process uses ammonia in the presence of a catalyst to selectively reduce NO

emissions from exhaust gases. In SCR, ammonia, usually diluted with air or steam, is injected through a grid system into hot flue gases that are then passed through a catalyst bed to carry out NO

reduction reactions. The catalyst is not consumed in the process but allows the reactions to occur at a lower temperature. However, SCR is subject to catalyst poisoning in high dust kiln exhausts. Therefore, SCR would have to be placed after the particulate control systems. According to LNA, given the operating temperature range for Kiln 1 and Kiln 2 at the Nelson Lime Plant, the SCR catalyst would need to be located prior to the kiln baghouses, which would result in poisoning or covering of the catalyst. In addition, there are no SCR systems currently operating on lime kilns. We propose to eliminate SCR as not technically feasible for retrofit on Kiln 1 and Kiln 2.

Selective Non-Catalytic Reduction:

SNCR is a technically feasible option for reducing NO

emissions from the Nelson Lime Plant kilns as shown in Table 20. This control technique relies on the reduction of NO

in exhaust gases by injection of ammonia or urea, without using any catalyst. This approach avoids the problems related to catalyst fouling and poisoning attributed to SCR, but requires injection of the reagents in the kiln at a temperature between 1600 °F to 2000 °F. Because no catalyst is used to increase the reaction rate, the temperature window is critical for conducting this reaction. LNA has not conducted any detailed design work for an SNCR system for the Nelson Plant kilns, but anticipates that a 50 percent reduction is achievable based on LNA's experience with operating a urea-injection system at another LNA lime plant.

Table 20—Nelson Lime Plant: SNCR Control Efficiency for Baseline Emissions

Control option

Control

efficiency

(%)

Emission

factor

(lb/ton lime)

Maximum emission rate

(lb/day)

(tpy)

Emissions

removed

(tpy)

Kiln 1:

Baseline

7.59

6,573

981

SNCR

3.80

3,286

491

Kiln 2:

Baseline

5.21

6,492

985

SNCR

2.61

3,246

493

b. BART Analysis for NO

EPA conducted a five-factor BART analysis of SNCR to evaluate its cost-effectiveness and visibility benefit. This analysis indicates that SNCR is cost-effective and results in visibility improvement.

Cost of Compliance:

The following table provides LNA's estimated cost for installation and operation of SNCR. Capital cost estimates developed by LNA relied primarily on vendor cost estimates and LNA's experience at other lime plants, with the remainder of the capital costs calculated using the cost methodology contained in EPA's Control Cost Manual. LNA has asserted a confidential business information (CBI) claim regarding certain annual operating costs such as reagent usage and auxiliary power costs. As a result, we have prepared our own independent estimate of annual operating costs based upon a combination of publicly available data and certain general assumptions as described in the Contractor's Report.

Table 21 is a summary of the estimated cost for installation and operation of SNCR.

Our estimate of annual operating costs is in the spreadsheet “Nelson Control Costs 2013-10-21.xlsx” in the docket.

Table 21—Nelson Lime Plant: Estimated Cost for SNCR

Kiln

Capital cost

($)

Annualized capital cost

($)

Annual

operating cost

($)

Total annual cost

($/yr)

Emission

reduction

(tpy)

Cost-

effectiveness

($/ton)

Kiln 1

$450,000

$42,477

$358,459

$400,936

491

$817

Kiln 2

450,000

42,477

354,981

397,458

493

807

Energy and Non-Air Quality Environmental Impacts:

SNCR systems require electricity to operate the blowers and pumps, which will likely involve fuel combustion that will generate emissions. Overall, while the generation of the required electricity will result in emissions, the emissions should be low compared to the reduction in NO

that would be gained by operating an SNCR system. The operation of SNCR systems on Kiln 1 and Kiln 2 would require that either urea or ammonia be stored on site. The storage of the chemicals does not result in a direct non-air quality impact. However, the potential for the urea or ammonia that would be stored to leak or otherwise be released from the storage vessels means there is the potential for both air and non-air quality related impacts. The storage of these chemicals does not significantly impact the BART determination.

Pollution Control Equipment in Use at the Source:

The presence of existing pollution control technology at each source is reflected in our BART analysis in two ways: first, in the consideration of available control technologies, and second, in the development of baseline emission rates for use in cost calculations and visibility modeling. Air pollution control equipment in use at the Nelson Lime Plant includes a number of baghouses, two multi-cyclone dust collectors, and a Ducon wet scrubber to control particulate matter emissions. The facility does not currently have control equipment for NO

and SO

The kilns are allowed to burn coal, petroleum coke, fuel oil, or any combination of these fuels.

Remaining Useful Life of the Source:

Since we are not aware of any enforceable shutdown date for the Nelson Lime Plant, we have used a 20-year amortization period, as noted in the EPA Cost Control Manual, as the remaining useful life of the kilns.

Degree of Visibility Improvement:

LNA performed a visibility analysis

to assess the visibility improvement associated with SNCR. LNA performed dispersion modeling using the CALPUFF modeling system, which consists of the CALPUFF dispersion model, the CALMET meteorological data processor, and the CALPOST post-processing program. The specific program versions that were relied upon in the analysis match the program versions relied upon by EPA's contractor, the University of North Carolina at Chapel Hill and ICF International (UNC/ICF), in the BART analyses that they prepared for select sources, including the Nelson Plant. Most of the same data and parameter settings relied upon in the analysis are the same data and parameter settings that were relied upon in the contractor's report. Compared to the UNC work, LNA used updated higher base case SO

and NOx emissions, lower PM emissions, and lower stack exit velocities. LNA's analysis included tables of visibility impacts and the improvement from controls, including results for the individual model years 2001, 2002, and 2003, and it used visibility method “8a” and focused on the highest value from among the three years' 98th percentiles. In order to put all the facilities on the same footing, EPA post-processed the modeling files provided by LNA using the approach followed for the other facilities.

BART Five Factor Analysis, Lhoist North America Nelson Lime Plant,

Trinity Consultants, August 2013.

Table 22 represents the 98th percentile by the 22nd high over the 2001-2003 period using visibility method “8b.” Using the EPA procedure, the maximum impact still occurs at the Grand Canyon, at 1.8 dv. The 98th percentile impacts at other Class I areas are about 0.3 dv or below, and the cumulative impact is 3.3 dv. The maximum visibility improvement due to SNCR is 0.58 dv, and cumulative improvement is 0.85 dv. There is little improvement at areas other than the Grand Canyon. These improvements yield a visibility cost-effectiveness of $1.4 million/dv using the maximum, and $0.9 million/dv using the cumulative improvement. These visibility improvements support the choice of SNCR as BART for NO

Table 22—Nelson Lime Plant: Visibility Impact and Improvement From NO

Controls

Class I area

Distance (km)

Visibility

impact

Base case

Visibility

improvement

SNCR

(ctr1)

Bryce Canyon NP

235

0.20

0.06

Grand Canyon NP

1.79

0.58

Joshua Tree NP

238

0.23

0.02

Mazatzal WA

206

0.15

0.01

Pine Mountain WA

199

0.15

0.02

Sierra Ancha WA

289

0.11

0.01

Superstition WA

288

0.13

0.01

Sycamore Canyon WA

132

0.31

0.07

Zion NP

183

0.28

0.08

Cumulative (sum)

3.34

0.85

Maximum

1.79

0.58

# CIAs >= 0.5 dv

Million $/dv (cumul. dv)

$0.9

Million $/dv (max. dv)

$1.4

c. Proposed BART Determination for NO

We propose to find that BART for NO

for Kilns 1 and 2 is SNCR, and are proposing a BART emission limit for Kiln 1 of 3.80 lb/ton lime and for Kiln 2 of 2.61 lb/ton lime on a 30-day rolling basis, as demonstrated through the use of a CEMS. We consider SNCR to be a very cost-effective control option for Kilns 1 and 2, at $817/ton and $807/ton, respectively. In addition, we consider the anticipated visibility benefit from SNCR, 0.58 dv at Grand Canyon National Park and 0.85 cumulatively at all Class I areas within 300 km, to be substantial. In considering the other factors, we do not consider their impact substantial relative to the cost and visibility factors. We note that the remaining useful life of the source is reflected in the evaluation of cost of compliance through the use of a 20-year amortization period in control cost calculations. Since there is no existing NO

control technology in use on the kilns, baseline emissions reflect uncontrolled NO

emissions. In examining energy and non-air quality impacts, while we note certain impacts associated with SNCR, we do not consider these impacts sufficient to warrant its elimination as a control option.

We propose to require compliance with this requirement within three years after the effective date of the final rule. A 2006 Institute of Clean Air Companies (ICAC) study indicated that the installation time for a typical SNCR retrofit, from bid to startup-up, is 10-13 months.

In relation to other industrial sources, such as fossil fuel boilers, there are a limited number of examples of SNCR installation on lime kilns. Given this relative lack of information regarding SNCR installation schedules on lime kilns, we consider three years to be an appropriate length of time to design, install, and test an ammonia injection system for a lime kiln. In addition, we are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this emission limit. As part of the proposed monitoring requirements, we are including a requirement to monitor rates of ammonia injection in order to ensure proper operation of the SNCR in a manner that minimizes ammonia emissions.

See “Typical Installation Timelines for NO

Emissions Control Technologies on Industrial Sources,” Institute of Clean Air Companies, December 4, 2006.

3. Proposed BART for SO

For our BART analysis, we identify all available control technologies, eliminate options that are not technically feasible, and evaluate the control effectiveness of the remaining control options. We then evaluate each control in terms of a five-factor BART analysis and make a determination for BART.

a. Control Technology Analysis for SO

EPA proposes to find that DSI and switching to lower sulfur fuel are technically feasible controls, while wet or semi-dry scrubbing is not technically feasible.

Wet or Semi-Dry Scrubbing:

We do not consider wet or dry scrubbing to be a feasible technology to control SO

emissions for this source. Wet scrubbing involves passing flue gas downstream from the main particulate matter control device through a sprayed aqueous suspension of lime or limestone that is contained in a scrubbing device. The SO

reacts with the scrubbing reagent to form lime sludge that is collected. The sludge usually is dewatered and disposed of at an offsite landfill. However, LNA has concluded, and we agree, that there is not sufficient water available for this type of system. According to LNA, two ground water wells supply about 106 gallons per minute (gpm) to the Nelson Plant, which currently uses about 80 gpm. Therefore, only 26 gpm of water is available for a scrubbing system that, even for a semi-dry scrubbing system that has lower water requirements than wet scrubbing, would require about 117 gpm. Moreover, a 1998 hydrologic report indicates that the prospects for developing additional wells, even low-yield wells, on the Nelson property are poor.

After reviewing the hydrologic report and the vendor estimate of water requirements for a semi-dry scrubber, we agree with this assessment.

See “Results of Hydrogeologic Investigations for Development of Additional Water Supply, Chemical Lime Company, Nelson Plant, Yavapai County, AZ,” July 8, 1998.

Dry Sorbent Injection:

DSI involves the injection of powdered absorbent directly into the flue gas exhaust stream. The sorbent reacts with SO

in the exhaust to form solid particles that are then removed by a particulate matter control device downstream of the sorbent injection. DSI is a simple system that generally requires a sorbent storage tank, feeding mechanism, transfer line and blower, and an injection device. DSI is generally considered technically feasible for the cement industry, although the level of control effectiveness may vary based upon site-specific conditions. We consider this option technically feasible for lime kilns. LNA has not included information in its analysis indicating

that DSI would be infeasible for the Nelson Plant kilns.

Lower Sulfur Fuel:

The lower sulfur fuel option described by LNA involves changing the proportion of coal and petroleum coke used as a fuel blend. LNA currently uses a blend of 27 percent coal and 73 percent petroleum coke, on a mass basis, as the fuel for the kilns. Since coke has about four to five times more sulfur than coal, it is possible to decrease the sulfur in the fuel blend by increasing the proportion of coal. However, an increase in coal in the fuel blend will also increase the ash content of the fuel blend. Ash in the fuel can disrupt operations due to the buildup of ash rings in the kilns. A fuel blend with an ash content of about 6.5 percent or less must be used in order to avoid these operational challenges.

As noted in fuel usage and purchase records, the Nelson Plant currently operates on a coal and petroleum coke mixture. As a result, we consider adjusting the coal/coke ratio in the fuel mixture to be a technically feasible option. We note, however, that since the BART Guidelines do not require fuel supply changes to be considered as a control option, we have typically not considered changes in fuel in BART analyses.

However, because LNA included lower sulfur fuel in its analysis, we have retained it as a control option.

40 CFR Part 51, Appendix Y, Section IV.D.1.5, “STEP 1: How do I identify all available retrofit emission control techniques?”

b. BART Analysis for SO

EPA conducted a five-factor BART analysis of the two technically feasible control options, DSI and lower sulfur fuel, to evaluate the cost-effectiveness and visibility benefit of each option along with any effect on the other factors.

Cost of Compliance:

Our consideration of the cost of compliance focuses primarily on the cost-effectiveness of each control option as measured in cost per ton and incremental cost per ton. We estimate the SO

emissions rates for DSI and lower sulfur fuel as shown in Table 23, and the cost-effectiveness of these options as shown in Table 24. DSI has a control efficiency of 40 percent that results in about 1,588 tpy of SO

removed from both kilns. Lower sulfur fuel has a control efficiency of 23.3 percent that results in about 925 tpy of SO

removed from both kilns. Based on the total annual costs of controlling SO

emissions at both kilns, DSI would cost an average of about $4,200 per ton removed and lower sulfur fuel about $860 per ton removed. Since there is no existing SO

control technology in use in the plant, baseline emissions reflect uncontrolled SO

emissions.

While we consider it appropriate to use 40 percent control efficiency

for DSI, we are inviting comment on the control effectiveness of 23.3 percent for a lower sulfur fuel blend based on the ratio of coal (1.15 percent sulfur) to petroleum coke (5.64 percent sulfur). LNA estimates that the maximum coal-to-coke ratio to maintain overall fuel ash content below 6.5 percent is a 50 percent coal to 50 percent coke fuel mixture. A 50/50 mix corresponds to a fuel sulfur reduction of 1.13 percentage points, which represents a 23.3 percent reduction from the current fuel mixture. Based on a review of coal and coke properties along with historical fuel usage at the Nelson Plant, we agree with the use of a 50/50 coal-to-coke ratio and 23.3 percent control effectiveness. However, LNA cites operational issues with fuel ash content above 6.5 percent. Since ash is a contaminant that can adversely affect lime product quality, we are seeking comment regarding the extent to which it is appropriate to use fuel ash content of 6.5 percent as the upper bound for determining fuel mixture ratio. We may finalize a different fuel mixture ratio based upon the comments we receive.

While the control efficiency for DSI is much higher for cement kilns, LNA conducted onsite testing of DSI on the lime kilns at the Nelson Plant that demonstrated it is appropriate to use 40 percent control efficiency. The docket includes a comparison of LNA's tests of DSI to the analysis in our contractor's report.

In estimating the costs of compliance, LNA relied on a vendor quote for purchased equipment provided by Noltech dated May 22, 2013, with the remainder of the capital costs calculated using the cost methodology contained in EPA's Control Cost Manual.

While these capital costs are higher than those estimated by our contractor, we consider the use of the Noltech vendor quote for the Nelson Plant reasonable, and have incorporated it into our evaluation of the costs of compliance. With regard to annual operating & maintenance costs, LNA has asserted a confidential business information (CBI) claim regarding certain annual operating costs such as reagent usage. As a result, we have prepared our own independent estimate of annual operating costs based upon a combination of publicly available data and certain assumptions as described in the contractor's report. Detailed cost calculations can be found in the docket.

Vendor quote included as an attachment to

BART Five Factor Analysis, Lhoist North America Nelson Lime Plant;

(Public version dated September 27, 2013).

See spreadsheet “Nelson Control Costs 2013-10-24.xlsx” in the docket.

Table 23—Nelson Lime Plant: SO

Control Option Emission Estimates

control technology

Control

efficiency

(%)

Emission factor

(lb/ton lime)

Maximum emission rate

lb/day

Tpy

Removed

(tpy)

Kiln 1:

Baseline

12.15

10,526

1,571

Lower Sulfur Fuel Blend

23.30

9.32

8,073

1,205

366

Dry Sorbent Injection

7.29

6,316

943

628

Kiln 2:

Baseline

12.69

15,808

2,400

Lower Sulfur Fuel Blend

23.30

9.73

12,125

1,841

559

Dry Sorbent Injection

7.61

9,485

1,440

960

Table 24—Nelson Lime Plant: SO

Control Option Cost-Effectiveness

control technology

Capital

cost

($)

Annual

direct costs

($/yr)

Annual

indirect costs

($/yr)

Total annual cost

($/yr)

Emission

reduction

(tpy)

Cost-effectiveness

($/ton)

Average

Incremental

Kiln 1:

Lower Sulfur Fuel Blend

$313,096

366

$856

Dry Sorbent Injection

$2,497,559

$371,174

$2,621,832

2,621,832

628

4,174

$8,803

Kiln 2:

Lower Sulfur Fuel Blend

458,179

559

819

Dry Sorbent Injection

2,497,559

371,174

3,895,774

960

4,058

8,576

Pollution Control Equipment in use at the Source:

The presence of existing pollution control technology at the Nelson Plant is reflected in the BART analysis in two ways: first, in the consideration of available control technologies, and second, in the development of baseline emission rates for use in cost calculations and visibility modeling. In the case of SO

the kilns at the Nelson Plant do not operate with any existing control technology. This is reflected in the baseline emission rates, which represent uncontrolled SO

emissions.

Energy and non-air quality environmental impacts:

Regarding the first option, DSI systems require electricity for operation. The generation of the electricity needed to operate a DSI system will likely involve fuel combustion that will generate emissions. Emissions also are associated with the transport, handling, and storage of sorbent. Overall, while the use of DSI will cause emissions from select activities, the emissions should be low compared to the reduction in SO

that would be gained by operating a DSI system. Regarding the second option, using a lower sulfur fuel blend means LNA will obtain more of the energy for lime production from coal and less of the energy from coke. Since the heating value of coke is slightly higher than the heating value of coal, it is likely that LNA will burn more total mass of fuel as a result of substituting some coal for coke. While burning a lower sulfur fuel blend will likely result in a reduction in SO

emissions, it will involve the overall use of greater quantities of coal, which may result in a collateral increase of other pollutants such as NO

and CO.

Remaining Useful Life of the Source:

We are considering the “remaining useful life” of the kilns as one element of the overall cost analysis as allowed by the BART Guidelines. In the absence of any enforceable closure date, we have used a 20-year amortization period described in the EPA Cost Control Manual as the remaining useful life for the control options considered for the Nelson Plant kilns. Since there is no capital costs associated with using a lower sulfur fuel blend, the remaining useful life of the kilns is not a factor in the evaluation of this technology.

Degree of Visibility Improvement:

As was the case for NO

EPA post-processed LNA's modeling results for SO

controls. The greatest improvement from DSI is 0.2 dv, occurring at the Grand Canyon, with improvements at other areas a third or less than this. The cumulative improvement is 0.6 dv. The maximum and cumulative improvements from switching to lower sulfur fuel are roughly half of these amounts. While visibility improvement by itself could support either DSI or lower sulfur fuel as BART, lower sulfur fuel is favored by its much lower average cost-effectiveness at $819-856/ton compared to over $4000 for DSI. Baseline and control option emission rates used in SO

control scenario modeling are summarized in Table 25 with the modeling results in Table 26.

These results are from EPA's post-processing of LNA's modeling. See the TSD for a discussion of the differences between EPA's results and the results reported by LNA in their BART analysis.

Table 25—Nelson Lime Plant: SO

Control Model Emission Rates

control technology

Control

efficiency

Emission factor

lb/ton lime

Maximum 24-hr

model emission rate

lb/day

lb/hr

g/s

Kiln 1:

Baseline

12.15

10,526

439

Lower Sulfur Fuel Blend

23.30

9.32

8,073

336

Dry Sorbent Injection (SBC)

7.29

6,315

263

Kiln 2:

Baseline

12.69

15,808

659

Lower Sulfur Fuel Blend

23.30

9.73

12,125

505

Dry Sorbent Injection (SBC)

7.61

9,489

395

Table 26—Nelson Lime Plant: SO

Control Option Visibility Modeling Results

Class I area

Distance (km)

Visibility

impact

Base case

Visibility

improvement

DSI

(ctr2)

Low-S fuel

(ctr3)

Bryce Canyon NP

235

0.20

0.03

0.02

Grand Canyon NP

1.79

0.21

0.10

Joshua Tree NP

238

0.23

0.07

0.04

Mazatzal WA

206

0.15

0.04

0.02

Pine Mountain WA

199

0.15

0.04

0.02

Sierra Ancha WA

289

0.11

0.04

0.02

Superstition WA

288

0.13

0.04

0.02

Sycamore Canyon WA

132

0.31

0.06

0.04

Zion NP

183

0.28

0.04

0.02

Cumulative (sum)

3.34

0.57

0.29

Maximum

1.79

0.21

0.10

# CIAs >= 0.5 dv

Million $/dv (cumul. dv)

$11.5

$2.6

Million $/dv (max. dv)

$30.7

$8.1

c. Proposed BART Determination for SO

We propose to find that BART for SO

is the use of a lower sulfur fuel blend with an emission limit of 9.32 lb/ton for Kiln 1 and 9.73 lb/ton for Kiln 2

on a rolling 30-day basis. In evaluating the costs of compliance, we note that we consider DSI and lower sulfur fuel to both be cost-effective control options, with average cost-effectiveness values of approximately $800/ton and $4,000/ton, respectively. In evaluating anticipated visibility benefit, while DSI is anticipated to achieve the greatest visibility improvement (0.21 dv at Grand Canyon), this amount of visibility improvement is not large, nor is the benefit anticipated for the next most stringent control option, lower sulfur fuel (0.10 dv at Grand Canyon). In considering the other factors, there is no significant effect on the outcome of the cost and visibility analyses. The lack of existing control technology is reflected in the baseline in the form of uncontrolled SO

emissions. In examining energy and non-air quality impacts, we note that there may be certain collateral increases in emissions, but that these increases are outweighed by the emission reductions achieved by implementing the control technology and do not warrant their elimination. The remaining useful life of the source is reflected in the evaluation of the cost of compliance. We consider both DSI and use of lower sulfur fuel to be cost-effective, but note that the most stringent option, DSI, is considerably less cost-effective than the use of lower sulfur fuel, with an incremental cost-effectiveness, relative to lower sulfur fuel, of approximately $9,000/ton. As a result, although DSI is the most stringent control option, the visibility benefit it achieves is not large, and is achieved at a very high incremental cost relative to the next most stringent control option. Based on this information, we propose to find that BART for SO

is the use of a lower sulfur fuel blend.

The differing emission limits are due to the different baseline performance of the two kilns.

4. Proposed BART for PM

For our BART analysis, we identified fabric filter baghouses, the existing control technology for PM

on Kilns 1 and 2, as the most stringent control available for this type of source.

a. Control Technology Analysis for PM

The Nelson Plant, as a major source of hazardous air pollutants (HAPs), is subject to the Maximum Achievable Control Technology (MACT) Standard for Lime Manufacturing Plants, and is required to meet an emission limit of 0.12 lbs PM/TSF (ton of stone feed).

The BART Guidelines provide that unless there are new technologies subsequent to the MACT standards that would lead to cost-effective increases in the level of control, one may rely on the MACT standards for purposes of BART.

Based on information developed as part of the Lime MACT, we estimate that existing fabric filter upgrades would result in annual costs of $94,500.

As noted in LNA's BART analysis, baseline PM emissions for the two kilns, based on PM filterable stack test data and annual lime production, are approximately 8 tpy and 15 tpy.

This would result in an average cost-effectiveness of about $6,300 to $12,000/ton.

40 CFR Part 63, Subpart AAAAA, Table 1, Item 1 for existing lime kilns with no wet scrubber prior to 2005.

40 CFR Part 51, Appendix Y, Section IV.C.

Annual costs as described in the Economic Impact Analysis for the Lime Manufacturing MACT Standard (EPA-452/R-03-013), Table 3-2, Model Kiln F. Adjusted from 1997 to 2013 dollars using the consumer price index, available at

ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt

As described in the LNA Nelson BART Analysis, Table 4-5.

b. BART Analysis for PM

The BART Guidelines provide that, in instances where a source already has the most stringent controls available (including all possible improvements), it is not necessary to complete each step of the BART analysis. Further, as long as the most stringent controls available are made federally enforceable for the purpose of implementing BART for that source, one may skip the remainder of the analysis, including the visibility analysis.

40 CFR Part 51, Appendix Y, Section IV.D.9.

c. Proposed BART Determination for PM

We propose a BART emission limit of 0.12 lb/TSF to control PM

at Kilns 1 and 2 based on the use of the existing fabric filter baghouses and commensurate with the MACT standard that applies to this source. We seek comment on any cost-effective upgrades or improvements that may result in a lower emission limit. We propose to require compliance with this requirement within 6 months after the effective date of the final rule. We also propose regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this emission limit that is found at the end of this notice.

C. Hayden Smelter

Summary:

EPA proposes to find that the ASARCO Hayden Smelter is subject to BART for NO

in addition to SO

determined by the State. ASARCO must capture and control SO

emissions from the converter units that are subject to BART. In the current method of operation, thousands of tons of SO

from these units are vented to the atmosphere with no pollution control. One method to control SO

emissions from the converter units is to install and operate a second double contact acid plant with a control efficiency of about 99.8 percent on a 30-day rolling average. We estimate the annual cost of constructing and operating a second acid plant to control SO

emissions is about $872 per ton of SO

removed. While we consider the cost of a new acid plant to be reasonable, we are proposing a performance standard as BART rather than prescribing a particular method of control. For NO

we propose to set an annual emission limit of 40 tpy from the BART-eligible units, based on our proposed determination that no NO

controls are needed for BART at the Hayden Smelter. Finally, we are proposing an emission limit and associated compliance requirements for PM

Affected Class I Areas:

Twelve Class I areas are within 300 km of the Hayden Smelter. Their nearest borders range from 48 km to 239 km away. Galiuro WA and Superstition WA are the closest, followed by Saguaro NP and Sierra Ancha WA. The highest baseline 98th percentile visibility impact is 1.7 dv at Superstition, with impacts at Galiuro slightly lower. Baseline visibility impacts at each of the twelve areas exceed 0.5 dv. The cumulative sum of visibility impacts over all the Class I areas is 12.1 dv.

Facility Overview:

ASARCO Hayden Smelter is a batch-process copper smelter in Gila County, Arizona. We previously approved ADEQ's determination that converters 1, 3, 4 and 5 and Anode Furnaces 1 and 2 at the facility are BART-eligible.

We also approved ADEQ's determination that these units are subject to BART for SO

and that BART for PM

at ASARCO Hayden is no additional controls. However, we disapproved ADEQ's determination that existing controls constitute BART for SO

and that the units are not subject to BART for NO

In light of these disapprovals and our FIP duty for regional haze in Arizona, we are required to promulgate a FIP to address BART for SO

and NO

78 FR 46412 (July 30, 2013). Please refer to the TSD for a description of these units.

Baseline Emissions Calculations:

Since neither ASARCO nor ADEQ identified baseline emissions for the Hayden Smelter, we calculated baseline emissions for SO

and NO

For SO

we used as the baseline the average of the two highest emitting years from the last five years that ASARCO reported to ADEQ. For NO

, we estimated emission rates based on the rated natural gas capacity of the burners in the four subject-to-BART converters and the two anode furnaces.

As indicated in Table 27, the majority of the source's SO

emissions (20,341 tpy of a total of 22,621 tpy) are process emissions from the converters. These process SO

emissions are collected through a secondary capture system, but are emitted uncontrolled through an annular stack that bypasses the existing double contact acid plant. While our BART analysis focuses on these uncontrolled SO

emissions from the converters, we also evaluated improved control of the SO

emissions from the existing acid plant and from the anode furnaces as well as controlling NO

emissions from all the BART units.

ASARCO Hayden Title V permit.

Table 27—Hayden Smelter: BART Baseline Emissions

[Tons per year]

Converters

Existing acid plant

(primary capture)

Annular stack

(secondary capture)

Uncaptured

Anode

furnaces

Total

1,034

20,341

1,209

22,621

Modeling Overview:

EPA is relying on modeled baseline and post-control impacts of the ASARCO Hayden Smelter using stack parameters provided by ASARCO in response to a 2013 EPA information request.

We also modeled using stack parameters based on a 2012 stack test.

Stack exit temperatures were comparable for these two models, but the exit velocities from the 2012 stack test were far lower than those provided by ASARCO in 2013. The 2012 stack test parameters resulted in visibility impacts and control benefits about 10 percent higher than the model using the 2013 parameters. We are conservatively using the 2013 ASARCO parameters to evaluate controls, since using the 2012 parameters would yield even greater visibility improvements. For both sets of modeling runs, EPA used emission rates that were developed using information provided by ASARCO.

Letter from Jack Garrity, ASARCO to Thomas Webb, EPA, July 11, 2013; attached Memorandum from Ralph Morris and Lynsey Parker, ENVIRON, to Eric Hiser, Jorden, Bischoff and Hiser, PLC, March 4, 2013.

ASARCO Hayden CEMS Test Report, Energy and Environmental Measurement Corporation, Test date: September, 2012.

1. BART Analysis and Determination for SO

From Converters

a. Control Technology Availability, Technical Feasibility and Effectiveness

EPA identified two available technology options to control the 20,341 tons of SO

emissions from the annular stack that are captured by a secondary collection system, but are released uncontrolled through the annular stack. These options are to construct and operate a second double contact acid plant or install a wet scrubber on the annular portion of the existing stack. In addition, we found that ASARCO could add a tail stack scrubber to the existing acid plant to address the remaining emissions that are not converted and removed as sulfuric acid by the acid plant. Regarding technical feasibility, we note that ASARCO Hayden currently uses a double contact acid plant to control SO

emissions captured by the primary capture system. Wet scrubbing also is commonly used in many industries to control SO

. Thus, we find

that the double contact acid plant and wet scrubbing are technically feasible. In terms of control effectiveness, ASARCO indicated in a letter

100

to EPA that its double contact acid plant is capable of recovering 99.8 percent of the SO

vented to it.

101

In the same letter, ASARCO noted that the expected control effectiveness of wet scrubbing is 85 percent. We used these removal efficiencies in our five-factor analyses. These analyses are explained in the TSD and summarized below.

100

Letter from Jack Garrity, ASARCO to Thomas Webb, EPA, July 11, 2013.

101

Ibid.

b. Option 1: Double Contact Acid Plant for Secondary Capture

Cost of Compliance:

EPA determined the cost-effectiveness of a new double contact acid plant is $872 per ton of SO

removed as shown in Table 28. As explained in the TSD, we conservatively estimated the cost of construction of a double contact acid plant to be $81,621,297. The annualized capital costs are based on a 20-year lifespan and a seven percent interest rate. We applied a control efficiency of 99.8 percent, which the existing acid plant is currently achieving with limited cesium catalyst. The emission reduction was applied to the secondary capture system baseline emissions. This cost analysis does not include the offsetting value of any sulfuric acid produced and sold. It does assume full catalyst replacement every other year and air preheating with natural gas for 8,760 hours per year.

Table 28—Hayden Smelter Option 1: Second Double Contact Acid Plant

Capital co

This text is long and has been trimmed here. Open the source document for the complete record.

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.