Standards of Performance for Petroleum Refineries

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 60

[EPA-HQ-OAR-2007-0011; FRL-8309-1]

RIN 2060-AN72

Standards of Performance for Petroleum Refineries

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Proposed rules.

SUMMARY:

EPA is proposing amendments to the current Standards of Performance for Petroleum Refineries. This action also proposes separate standards of performance for new, modified, or reconstructed process units at petroleum refineries. Unless otherwise noted, the term new includes modified or reconstructed units. The proposed standards for new process units include emissions limitations and work practice standards for fluid catalytic cracking units, fluid coking units, delayed coking units, process heaters and other fuel gas combustion devices, fuel gas producing units, and sulfur recovery plants. These proposed standards reflect demonstrated improvements in emissions control technologies and work practices that have occurred since promulgation of the current standards.

DATES:

Comments.

Written comments must be received on or before July 13, 2007.

Public Hearing.

If anyone contacts EPA by June 4, 2007 requesting to speak at a public hearing, a public hearing will be held on June 13, 2007.

ADDRESSES:

Submit your comments, identified by Docket ID No. EPA-HQ-OAR-2007-0011, by one of the following methods:

•

http://www.regulations.gov:

Follow the on-line instructions for submitting comments.

•

E-mail: a-and-r-docket@epa.gov

.

•

Fax:

(202) 566-1741.

•

Mail: U.S. Postal Service, send comments to:

EPA Docket Center (6102T), New Source Performance Standards for Petroleum Refineries Docket, 1200 Pennsylvania Ave., NW., Washington, DC 20460. Please include a total of two copies. In addition, please mail a copy of your comments on the information collection provisions to the Office of Information and Regulatory Affairs, Office of Management and Budget (OMB),

Attn:

Desk Officer for EPA, 725 17th St., NW., Washington, DC 20503.

•

Hand Delivery: In person or by courier, deliver comments to:

EPA Docket Center (6102T), New Source Performance Standards for Petroleum Refineries Docket, EPA West, Room 3334, 1301 Constitution Avenue, NW., Washington, DC 20004. Such deliveries are only accepted during the Docket's normal hours of operation, and special arrangements should be made for deliveries of boxed information. Please include a total of two copies.

Instructions:

Direct your comments to Docket ID No. EPA-HQ-OAR-2007-0011. EPA's policy is that all comments received will be included in the public docket without change and may be made available online at

http://www.regulations.gov

, including any personal information provided, unless the comment includes information claimed to be Confidential Business Information (CBI) or other information whose disclosure is restricted by statute. Do not submit information that you consider to be CBI or otherwise protected through

http://www.regulations.gov

or e-mail. The

http://www.regulations.gov

website 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 e-mail comment directly to EPA without going through

http://www.regulations.gov

, your e-mail address will be automatically captured and included 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 avoid the use of special characters, any form of encryption, and be free of any defects or viruses.

Docket:

All documents in the docket are listed in the

http://www.regulations.gov

index. Although listed in the index, some information is not publicly available, e.g., CBI or other information whose disclosure is restricted by statute. Certain other material, such as copyrighted material, will be publicly available only in hard copy. Publicly available docket materials are available either electronically in

http://www.regulations.gov

or in hard copy at the EPA Docket Center, Standards of Performance for Petroleum Refineries Docket, EPA West, Room 3334, 1301 Constitution Ave., NW., Washington, DC. The Public Reading Room is open from 8:30 a.m. to 4:30 p.m., Monday through Friday, excluding legal holidays. The telephone number for the Public Reading Room is (202) 566-1744, and the telephone number for the Docket Center is (202) 566-1742.

FOR FURTHER INFORMATION CONTACT:

Mr. Robert B. Lucas, Office of Air Quality Planning and Standards, Sector Policies and Programs Division, Coatings and Chemicals Group (E143-01), Environmental Protection Agency, Research Triangle Park, NC 27711, telephone number: (919) 541-0884;

fax number:

(919) 541-0246;

e-mail address: lucas.bob@epa.gov

.

SUPPLEMENTARY INFORMATION:

I. General Information

A. Does this action apply to me?

Categories and entities potentially regulated by this proposed rule include:

Category

NAICS code

1

Examples of regulated entities

Industry

32411

Petroleum refiners.

Federal government

Not affected.

State/local/tribal government

Not affected.

1

North American Industrial Classification System.

This table is not intended to be exhaustive, but rather provides a guide for readers regarding entities likely to be regulated by this action. To determine whether your facility would be regulated by this action, you should examine the applicability criteria in 40 CFR 60.100 and 40 CFR 60.100a. If you have any questions regarding the applicability of this proposed action to a particular entity, contact the person listed in the preceding

FOR FURTHER INFORMATION CONTACT

section.

B. What should I consider as I prepare my comments to EPA?

Do not submit information containing CBI to EPA through

http://www.regulations.gov

or e-mail. Send or deliver information identified as CBI only to the following address: Roberto Morales, OAQPS Document Control Officer (C404-02), Office of Air Quality Planning and Standards, Environmental Protection Agency, Research Triangle Park, NC 27711, Attention Docket ID No. EPA-HQ-OAR-2007-0011. Clearly mark the part or all of the information that you claim to be 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 then 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, a copy of the comment that does not contain the information claimed as CBI must be submitted for inclusion in the public docket. Information so marked will not be disclosed except in accordance with procedures set forth in 40 CFR part 2.

C. Where can I get a copy of this document?

In addition to being available in the docket, an electronic copy of this proposed action is available on the Worldwide Web (WWW) through the Technology Transfer Network (TTN). Following signature, a copy of this proposed action will be posted on the TTN's policy and guidance page for newly proposed or promulgated rules at

http://www.epa.gov/ttn/oarpg

. The TTN provides information and technology exchange in various areas of air pollution control.

D. When would a public hearing occur?

If anyone contacts EPA requesting to speak at a public hearing by June 4, 2007, a public hearing will be held on June 13, 2007. Persons interested in presenting oral testimony or inquiring as to whether a public hearing is to be held should contact Mr. Bob Lucas, listed in the

FOR FURTHER INFORMATION CONTACT

section, at least 2 days in advance of the hearing.

E. How is this document organized?

The supplementary information presented in this preamble is organized as follows:

I. General Information

A. Does this action apply to me?

B. What should I consider as I prepare my comments to EPA?

C. Where can I get a copy of this document?

D. When would a public hearing occur?

E. How is this document organized?

II. Background Information

A. What is the statutory authority for the proposed standards and proposed amendments?

B. What are the current petroleum refinery NSPS?

III. Summary of the Proposed Standards and Proposed Amendments

A. What are the proposed amendments to the standards for petroleum refineries (40 CFR part 60, subpart J)?

B. What are the proposed requirements for new fluid catalytic cracking units and new fluid coking units (40 CFR part 60, subpart Ja)?

C. What are the proposed requirements for new sulfur recovery plants (SRP) (40 CFR part 60, subpart Ja)?

D. What are the proposed requirements for new process heaters and other fuel gas combustion devices (40 CFR part 60, subpart Ja)?

E. What are the proposed work practice and equipment standards (40 CFR part 60, subpart Ja)?

IV. Rationale for the Proposed Amendments (40 CFR part 60, subpart J)

A. How is EPA proposing to change requirements for refinery fuel gas?

B. How is EPA proposing to amend definitions?

C. How is EPA proposing to revise the coke burn-off equation?

D. What miscellaneous corrections are being proposed?

V. Rationale for the Proposed Standards (40 CFR part 60, subpart Ja)

A. What is the performance of control technologies for fluid catalytic cracking units?

B. What is the performance of control technologies for fuel gas combustion?

C. What is the performance of control technologies for process heaters?

D. What is the performance of control technologies for sulfur recovery systems?

E. How did EPA determine the proposed standards for new petroleum refining process units?

VI. Modification and Reconstruction Provisions

VII. Request for Comments

VIII. Summary of Cost, Environmental, Energy, and Economic Impacts

A. What are the impacts for petroleum refineries?

B. What are the secondary impacts?

C. What are the economic impacts?

D. What are the benefits?

IX. 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 Concerning Regulations That Significantly Affect Energy Supply, Distribution, or Use

I. National Technology Transfer Advancement Act

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

II. Background Information

A. What is the statutory authority for the proposed standards and proposed amendments?

New source performance standards (NSPS) implement Clean Air Act (CAA) section 111(b) and are issued for categories of sources which cause, or contribute significantly to, air pollution which may reasonably be anticipated to endanger public health or welfare. The primary purpose of the NSPS is to attain and maintain ambient air quality by ensuring that the best demonstrated emission control technologies are installed as the industrial infrastructure is modernized. Since 1970, the NSPS have been successful in achieving long-term emissions reductions in numerous industries by assuring cost-effective controls are installed on new, reconstructed, or modified sources.

Section 111 of the CAA requires that NSPS reflect the application of the best system of emission reductions which (taking into consideration the cost of achieving such emission reductions, any non-air quality health and environmental impact and energy requirements) the Administrator determines has been adequately demonstrated. This level of control is commonly referred to as best demonstrated technology (BDT).

Section 111(b)(1)(B) of the CAA requires EPA to periodically review and revise the standards of performance, as necessary, to reflect improvements in methods for reducing emissions.

B. What are the current petroleum refinery NSPS?

NSPS for petroleum refiners (40 CFR part 60, subpart J) apply to fluid catalytic cracking unit catalyst regenerators and fuel gas combustion devices that commence construction or modification after June 11, 1973. Fluid catalytic cracking unit catalyst regenerators are subject to standards for particulate matter (PM), opacity, and carbon monoxide (CO). Fluid catalytic cracking unit catalyst regenerators that commence construction after January 17, 1984 are also subject to standards for sulfur dioxide (SO

2

) (or a feed sulfur content limit). Fuel gas combustion devices are subject to concentration limits for hydrogen sulfide (H

2

S) as a surrogate for SO

2

emissions.

The current NSPS also apply to all Claus sulfur recovery plants (SRP) of more than 20 long tons per day (LTD) that commence construction or modification after October 4, 1976. Claus SRP are subject to standards for either SO

2

or both reduced sulfur compounds and H

2

S.

The NSPS were originally promulgated on March 8, 1974 and have been amended several times. Significant changes to emission limits since the original promulgation date include the addition of the sulfur oxide standards for SRP and fluid catalytic cracking units (see 43 FR 10869, March 15, 1978 and 54 FR 34027, August 17 1989).

III. Summary of the Proposed Standards and Proposed Amendments

We are proposing several amendments to provisions in the existing NSPS in 40 CFR part 60, subpart J. Many of these amendments are technical clarifications and corrections that are also included in the proposed standards in 40 CFR part 60, subpart Ja. For example, we are proposing language to change the definition of fuel gas to indicate that vapors collected and combusted to comply with certain wastewater and marine vessel loading provisions are not considered fuel gas and are exempt from 40 CFR 60.104(a)(1). These gas streams are not required to be monitored. In a related amendment, we are proposing to clarify that monitoring is not required for fuel gases that are identified as inherently low sulfur or can demonstrate a low sulfur content. We are also revising the coke burn-off equation to account for oxygen (O

2

)-enriched air streams. Other amendments include clarification of definitions and correction of grammatical and typographical errors.

The proposed standards in 40 CFR part 60, subpart Ja include emission limits for fluid catalytic cracking units, fluid coking units, SRP, and fuel gas combustion devices. They also include work practice standards for minimizing the quantity of fuel gas streams flared from all refinery process units and for minimizing the SO

2

emissions from process units that are subject to standards of performance for SO

2

emissions. Proposed equipment standards would reduce emissions of volatile organic compounds (VOC) from delayed coker units. Only those affected facilities that begin construction, modification, or reconstruction after May 14, 2007 would be affected by the proposed standards in 40 CFR part 60, subpart Ja. Units for which construction, modification, or reconstruction began on or before May 14, 2007 would continue to comply with the applicable standards under the current NSPS in 40 CFR part 60, subpart J, as amended.

A. What are the proposed amendments to the standards for petroleum refineries (40 CFR part 60, subpart J)?

We are proposing to amend the definition of “fuel gas” to exempt vapors that are collected and combusted in an air pollution control device installed to comply with a specified wastewater or marine vessel loading emissions standard. The thermal combustion control devices themselves would still be considered affected fuel gas combustion devices, and all auxiliary fuel fired to these devices would be subject to the fuel gas limit; however, continuous monitoring would not be required for the collected vapors that are being incinerated because these gases would not be considered fuel gases under the proposed definition of “fuel gas” in subpart J.

We are also proposing to exempt certain fuel gas streams from all continuous monitoring requirements. Monitoring is currently not required for events that are exempt from the requirements in 40 CFR 60.104(a)(1) (flaring of process upset gases or flaring of gases from relief valve leakage or emergency malfunctions). Additionally, monitoring would not be required for inherently low sulfur fuel gas streams. These streams include pilot gas flames, gas streams that meet commercial-grade product specifications with a sulfur content 30 parts per million by volume (ppmv) or less, fuel gases produced by process units that are intolerant to sulfur contamination, and fuel gas streams that an owner or operator can demonstrate are inherently low-sulfur. Owners and operators would be required to document the exemption for which each fuel gas stream applies and ensure that the stream remains qualified for that exemption.

We are proposing to amend the definitions of “Claus sulfur recovery plant,” “oxidation control system,” and “reduction control system” to clarify that a SRP may consist of multiple units, that sulfur pits are part of the Claus SRP, and that the oxidized or reduced sulfur is recycled to the beginning of a sulfur recovery train within the SRP. We are also proposing to add a fourth term to the coke burn-off rate equation to account for the use of O

2

-enriched air.

Finally, the proposed amendments include a few technical corrections to fix references and other miscellaneous errors in subpart J. The specific changes are detailed in section IV.D of this preamble.

B. What are the proposed requirements for new fluid catalytic cracking units and new fluid coking units (40 CFR part 60, subpart Ja)?

The proposed standards for new fluid catalytic cracking units include emission limits for PM, SO

2

, nitrogen oxides (NO

X

), and CO. One difference from the existing standards in subpart J is that new fluid coking units would be subject to the same standards as fluid catalytic cracking units. Other differences from the existing standards are that the proposed PM and SO

2

emission limits are more stringent and the NO

X

emission limit is a new requirement. Unlike the existing standards, the proposed standards include no opacity limit because the opacity limit was intended to ensure compliance with the PM limit and because we are now proposing that sources use direct PM monitoring or parameter monitoring to ensure compliance with the PM limit.

The proposed PM emission limit for new fluid catalytic cracking units and new fluid coking units is 0.5 kilogram (kg) per Megagram (kg/Mg) (0.5 pound (lb)/1,000 lb) of coke burn-off in the regenerator. Initial compliance with this emission limit would be determined using Method 5 in Appendix A to 40 CFR part 60. Procedures for computing the PM emission rate using the total PM concentration, effluent gas flow rate, and coke burn-off rate would be the same as in 40 CFR part 60, subpart J, as amended. To demonstrate ongoing compliance, an owner or operator must either monitor PM emission control device operating parameters or use a PM continuous emission monitoring system (CEMS). If operating parameters will be used to demonstrate ongoing compliance, the owner or operator must monitor the same parameters during the initial performance test, and develop operating parameter limits for the applicable parameters. The operating limits must be based on the lowest hourly average values for the applicable parameters measured over the three test runs. The owner or operator must also conduct additional performance tests at least once every 24 months to verify compliance with the PM emission limit and confirm or reestablish operating limits. If ongoing compliance will be demonstrated using a PM CEMS, the CEMS must meet the conditions in Performance Specification 11. Thus, separate performance tests are not required because the equivalent of an initial performance test will be part of the initial correlation test for the PM CEMS, and periodic response correlation audits (every 5 years) will

include the equivalent of performance tests. We are co-proposing requiring reconstructed and modified fluid catalytic cracking units to meet the current standards in 40 CFR part 60, subpart J, and we are requesting comments on the effects of the proposed PM standard on modified or reconstructed facilities and if it is appropriate to adopt a different standard for these sources.

The proposed SO

2

emission limits for new fluid catalytic cracking units and new fluid coking units are to maintain SO

2

emissions to the atmosphere less than or equal to 50 ppmv on a 7-day rolling average basis, and less than or equal to 25 ppmv on a 365-day rolling average basis (both limits corrected to 0 percent moisture and 0 percent excess air). Initial compliance with the proposed 50 ppmv SO

2

emission limit would be demonstrated by conducting a performance evaluation of the SO

2

CEMS in accordance with Performance Specification 2 in appendix B of 40 CFR part 60, with Method 6, 6A, or 6C of 40 CFR part 60, appendix A as the reference method. Ongoing compliance with both proposed SO

2

emission limits would be determined using the CEMS to measure SO

2

emissions as discharged to the atmosphere, averaged over the 7-day and 365-day averaging periods. Rolling average concentrations would be calculated once per day using the applicable number of daily average values. We are co-proposing requiring reconstructed and modified fluid catalytic cracking units to meet the current standards in 40 CFR part 60, subpart J, and we are requesting comments on the effects of the proposed SO

2

standard on modified or reconstructed facilities.

The proposed NO

X

emission limits for new fluid catalytic cracking units and new fluid coking units are 80 ppmv on a 7-day rolling average basis (dry at 0 percent excess air). Initial compliance with the 80 ppmv emission limit would be demonstrated by conducting a performance evaluation of the CEMS in accordance with Performance Specification 2 in appendix B to 40 CFR part 60, with Method 7 of 40 CFR part 60, subpart A as the Reference Method. Ongoing compliance with this emission limit would be determined using the CEMS to measure NO

X

emissions as discharged to the atmosphere, averaged over 7-day periods. We are also co-proposing no new standards for NO

X

emissions from fluid coking units and for modified or reconstructed fluid catalytic cracking units.

The proposed CO emission limit for new fluid catalytic cracking units and new fluid coking units is 500 ppmv (1-hour average, dry at 0 percent excess air). Initial compliance with this emission limit would be demonstrated by conducting a performance evaluation for the CEMS in accordance with Performance Specification 4 in appendix B to 40 CFR part 60, with Method 10 or 10A in 40 CFR part 60, appendix A as the Reference Method. For Method 10, the integrated sampling technique is to be used. Ongoing compliance with this emission limit would be determined on an hourly basis using the CEMS to measure CO emissions as discharged to the atmosphere. An exemption from monitoring may be requested if the owner or operator can demonstrate that average CO emissions are less than 50 ppmv (dry basis). This limit and the compliance procedures are the same as in the existing NSPS for fluid catalytic cracking units.

C. What are the proposed requirements for new sulfur recovery plants (SRP) (40 CFR part 60, subpart Ja)?

The proposed standards include SO

2

emission limits for all SRP. The proposed emission limit for new SRP greater than 20 LTD is 250 ppmv or less of combined SO

2

and reduced sulfur compounds as discharged to the atmosphere (reported as SO

2

on a dry basis at 0 percent excess air). For a SRP with a capacity of 20 LTD or less, the proposed standard is mass emissions of combined SO

2

and reduced sulfur compounds equal to 1 weight percent or less of sulfur recovered. In addition, the proposed standards include an H

2

S concentration limit of 10 ppmv or less (dry basis at 0 percent excess air) for all new SRP. Both SO

2

and H

2

S concentration limits would be determined hourly on a 12-hour rolling average basis. As in the amendments to subpart J, the proposed definition of a SRP would include the sulfur pit.

Initial compliance with the emission limit for combined SO

2

and reduced sulfur compounds is demonstrated by conducting a performance evaluation for the SO

2

CEMS in accordance with Performance Specification 2 in appendix B to 40 CFR part 60, with Method 6, 6A, or 6C in 40 CFR part 60, appendix A as the Reference Method to determine the SO

2

concentration, and Method 15 in 40 CFR part 60, appendix A as the Reference Method to determine the SO

2

-equivalent concentration of the reduced sulfur compounds. The results of the test using Method 15 are also used to demonstrate initial compliance with the H

2

S concentration limit. Initial compliance with the mass sulfur emission limit is demonstrated by conducting a performance test as described above to determine the combined SO

2

and SO

2

-equivalent concentration, and then converting that concentration to a mass fraction using the volumetric flow rate of effluent gas and the mass rate of sulfur recovery during the performance test.

Ongoing compliance with the combined SO

2

and reduced sulfur compounds emission limit would be determined using a CEMS that uses an air or O

2

dilution and oxidation system to convert the reduced sulfur to SO

2

and then measures the total resultant SO

2

concentration. An O

2

monitor would also be required for converting the measured combined SO

2

concentration to the concentration at 0 percent O

2

. Ongoing compliance with the mass sulfur emission limit would be determined using the same types of CEMS. A flow monitor that continuously monitors the volumetric flow rate of gases released to the atmosphere would be required so that the mass emitted can be calculated. The hourly sulfur production rates would also have to be tracked so that mass fraction emitted can be calculated and compared with the proposed 1 percent emission limit.

Ongoing compliance with the H

2

S concentration limit would be determined using either an H

2

S CEMS or, if the SRP is equipped with an oxidation control system or followed by incineration, by continuous monitoring of the operating temperature and O

2

concentration. Minimum operating limits for the operating temperature and O

2

concentration would be established during the performance test.

D. What are the proposed requirements for new process heaters and other fuel gas combustion devices (40 CFR part 60, subpart Ja)?

The proposed standards for new process heaters include both SO

2

and NO

X

emission limits. Because of this, the fuel gas combustion units as defined in the existing subpart J standards were divided into two separate affected sources: “process heaters” and “other fuel gas combustion devices.” The primary sulfur oxides emission limit for new process heaters and other fuel gas combustion devices is 20 ppmv or less SO

2

(dry at 0 percent excess air) on a 3-hour rolling average basis and 8 ppmv or less on a 365-day rolling average basis. For process heaters that use only fuel gas and other fuel gas combustion devices, we are proposing an alternative concentration limit of 160 ppmv or less H

2

S or total reduced sulfur (TRS) in the fuel gas on a 3-hour rolling average basis (as in the existing NSPS) and 60 ppmv or less H

2

S or TRS in the fuel gas on a

365-day rolling averaging basis. The TRS concentration limit is required for new fuel gas combustion devices that combust fuel gas generated from coking units (as either the only fuel or as a mixture of fuel gases from other units). On the other hand, new fuel gas combustion devices that do not combust fuel gas generated from coking units are required to monitor H

2

S concentrations. Compliance would be demonstrated either by measuring H

2

S (or TRS) in the fuel gas or by measuring SO

2

in the exhaust gas.

Initial compliance with the 20 ppmv SO

2

limit or the 160 ppmv H

2

S or TRS concentration limits would be demonstrated by conducting a performance evaluation for the CEMS. The performance evaluation for an SO

2

CEMS would be conducted in accordance with Performance Specification 2 in appendix B to 40 CFR part 60, with Method 6, 6A, or 6C as the Reference Method. The performance evaluation for an H

2

S CEMS would be conducted in accordance with Performance Specification 7 in 40 CFR part 60, with Method 11, 15, 15A, or 16 as the Reference Method. The performance evaluation for a TRS CEMS would be conducted in accordance with Performance Specification 7 in 40 CFR part 60, with Method 16 as the Reference Method. Ongoing compliance with the proposed sulfur oxides emission limits would be determined using the applicable CEMS to measure either H

2

S or TRS in the fuel gas being used for combustion or SO

2

in the exhaust gas to the atmosphere, averaged over the 3-hour and 365-day averaging periods.

Similar to proposed clarifications for 40 CFR part 60, subpart J, we are proposing a definition of “fuel gas” that includes exemptions for vapors collected and combusted in an air pollution control device installed to comply with specified wastewater or marine vessel loading provisions. Also similar to subpart J, we are proposing to exempt from continuous monitoring fuel gas streams exempt under 40 CFR 60.102a(i) and fuel gas streams that are inherently low in sulfur. We are also proposing to streamline the process for an owner or operator to demonstrate that a fuel gas stream not explicitly exempted from continuous monitoring is inherently low sulfur.

The proposed NO

X

emission limits for new process heaters is 80 ppmv on a 7-day rolling average basis (dry at 0 percent excess air). Initial compliance with the 80 ppmv emission limit would be demonstrated by conducting a performance evaluation of the CEMS in accordance with Performance Specification 2 in appendix B to 40 CFR part 60, with Method 7 of 40 CFR part 60, subpart A as the Reference Method. Ongoing compliance with this emission limit would be determined using the CEMS to measure NO

X

emissions as discharged to the atmosphere, averaged over 7-day periods.

E. What are the proposed work practice and equipment standards (40 CFR part 60, subpart Ja)?

Three work practice standards are proposed to reduce both VOC and SO

2

emissions from flares, start-up/shutdown/malfunction events, and delayed coker units. First, the proposed rule requires all new fuel gas producing units at a refinery to be designed and operated in such a way that the fuel gas produced by the new process units does not routinely discharge to a flare. Second, a requirement for a start-up, shutdown and malfunction plan that includes procedures to minimize discharges either directly to the atmosphere or to the flare gas system during the planned startup or shutdown of these units, procedures to minimize emissions during malfunctions of the amine treatment system or sulfur recovery plant, and procedures for conducting a root-cause analysis of an emissions limit exceedance or process start-up, shutdown, upset, or malfunction that causes a discharge into the atmosphere, either directly or indirectly, from any refinery process unit subject to the provisions of this subpart in excess of 500 lb per day (lb/d) of SO

2

. Third, the proposed rule would require delayed coking units to depressure to 5 lbs per square inch gauge (psig) during reactor vessel depressuring and vent the exhaust gases to the fuel gas system. For new, reconstructed, or modified units, we are co-proposing to require only the last of these work practice standards, the requirement to depressure coking units to the flare.

IV. Rationale for the Proposed Amendments (40 CFR part 60, subpart J)

Because we are proposing a new subpart to 40 CFR part 60 for affected sources at petroleum refineries beginning construction, reconstruction, or modification after May 14, 2007, our proposed amendments to subpart J of 40 CFR part 60 would impact only those affected sources that are already subject to 40 CFR part 60, subpart J. The proposed amendments to this subpart include clarifications of the current requirements and technical corrections to the regulatory language. These changes to subpart J of 40 CFR part 60 are discussed below.

A. How is EPA proposing to change requirements for refinery fuel gas?

As we conducted our review of 40 CFR part 60, subpart J, we found that the definition of “fuel gas” has been broadly interpreted by States and EPA Regions over the last 30 years. Because of the increasing complexity of petroleum refineries, this interpretation may be more inclusive than originally intended in the 1970s. We agree that the interpretation ensures that all streams that could be considered fuel gas and have the potential for high-sulfur emissions are included in the regulatory requirements, but we recognize that this broad definition has resulted in application of the fuel gas concentration limits to fuel gas streams and combustion devices that were not originally considered in the standards development process. Furthermore, had these extended applications been considered in the standards development process, some of the applications would have been found to be either technically or economically infeasible. The existing requirements in subpart J of 40 CFR part 60 do recognize and limit the applicability of the fuel gas concentration limits to certain gas streams. For example, 40 CFR 60.101(d) excludes gases generated by catalytic cracking unit catalyst regenerators and fluid coking burners from the definition of “fuel gas.” These gases were excluded because the sulfur in the gases generated by the catalytic cracking unit catalyst regenerators and fluid coking burners is in the form of sulfur oxides rather than H

2

S. As such, these gases are not amenable to amine treatment, which was the primary treatment technique on which the fuel gas concentration limits were based. In addition, 40 CFR 60.104(a)(1) exempts process upset gases or fuel gas released to the flare as a result of relief valve leakage or emergency malfunctions from the fuel gas H

2

S concentration limits. In this case, it was determined that requiring treatment of these gases was either technically or economically infeasible. Therefore, it is entirely in keeping with the regulatory intent of the NSPS and the specific requirements in 40 CFR part 60, subpart J to exclude or exempt sources based on technical and economic considerations.

Since the development of the refinery fuel gas concentration limits in the early 1970s, EPA has developed numerous other standards in which incineration was promoted as a best air pollution management practice for certain organic vapors which had traditionally been

released directly to the atmosphere. These gas streams were never considered in the development of the 40 CFR part 60, subpart J standards because they were not directed to a fuel gas combustion device at the time. As such, the technical and economical feasibility of meeting the fuel gas concentration limits was not specifically evaluated for these gas streams at that time. During our review, we evaluated the application of the fuel gas concentration limits to a variety of process gas streams that did not exist in the early 1970s. We concluded that most of these gas streams are amenable to amine treatment and that it is both technically and economically feasible to treat those gas streams to meet the fuel gas concentration limits. However, we identified a few specific streams that are not readily amenable to amine treatment (or direct diversion to the SRP) and/or are not cost-effective to amine treatment due to the typically low (but potentially variable) H

2

S content and the typical location of these gas streams in relationship to the primary processing units at the refinery.

As a result of this evaluation, we are proposing to change the requirements of the fuel gas concentration limits in keeping with a broad definition of fuel gas, but recognizing the technical and economic issues related to certain fuel gas streams or combustion devices. Specifically, we are proposing to exempt from the definition of “fuel gas” vapors that are collected and combusted in an air pollution control device installed to comply with the Standards of Performance for VOC Emissions From Petroleum Refinery Wastewater Systems (40 CFR part 60, subpart QQQ), National Emission Standards for Benzene Waste Operations (40 CFR part 61, subpart FF), the National Emission Standards for Marine Tank Vessel Loading Operations (40 CFR part 63, subpart Y), or the National Emission Standards for Hazardous Air Pollutants From Petroleum Refineries (40 CFR part 63, subpart CC), specifically either 40 CFR 63.647 or 40 CFR 63.651. The wastewater and marine vessel loading sources subject to these specific regulations are often located at the edge of the refinery property, if not off-site, and compliance with the regulations is generally demonstrated by capturing and combusting the organic vapors. The collected gases generally have low sulfur content, but variability in the products being loaded and in wastewater treatment process operations may result in the collected gases exceeding the current fuel gas concentration limits for short periods of time. Due to the typical low sulfur content of these gases, they are not generally suitable for amine treatment; due to the presence of O

2

in these collected gases, they cannot be routed to the fuel gas system. Furthermore, these sources are typically far from amine treatment or the SRP, and it is not economically reasonable to propose control beyond the existing regulations for these sources (e.g., requiring these streams to be routed to sulfur treatment rather than being combusted). Therefore, we are proposing to amend the definition of “fuel gas” in 40 CFR 60.101(d) to exclude from the fuel gas concentration limits the vapors collected and combusted in air pollution control devices to comply with the specified regulations in 40 CFR part 60, subpart QQQ, 40 CFR part 61, subpart FF, or 40 CFR part 63, subparts Y or CC. The thermal combustion control devices would still be considered affected fuel gas combustion devices and all auxiliary fuel fired to these devices would be subject to the fuel gas concentration limit; however, continuous monitoring would not be required for the collected vapors that are being incinerated because these gases would not be considered fuel gases under the proposed definition of “fuel gas” in subpart J.

We are also proposing to clarify that monitoring is not required for fuel gas streams that are exempt from the requirements in 40 CFR 60.104(a)(1). These streams include process upset gases or fuel gases that are released to the flare as a result of relief valve leakage or other emergency malfunctions. To clarify this point, the proposed introductory text for 40 CFR 60.105(a)(4)(iv) specifies that continuous monitoring is not required for streams that are exempt from 40 CFR 60.104(a)(1). We are also proposing to add the phrase “for fuel gas combustion devices subject to 40 CFR 60.104(a)(1)” after “Instead of the SO

2

monitor in paragraph (a)(3) of this section” in 40 CFR 60.105(a)(4). This proposed amendment is more consistent with the language in 40 CFR 60.105(a)(3). Given our intent not to require fuel gas monitoring of process upset gases, combustion devices such as emergency flares would likely not require monitoring unless sources other than process upset gases are burned, such as routine vents or sweep gas. We are aware of issues related to the identification and exemption of these units from fuel gas monitoring. We are requesting comment on the need to provide specific language exempting these units, and on appropriate methods for identifying emergency flares and verifying on an ongoing basis that no flaring of nonexempt gases is occurring.

In addition to the exemptions described in the previous paragraphs, we are proposing to exempt certain fuel gas streams from all monitoring requirements. These streams would still be subject to the fuel gas concentration limits, but since we do not expect that these streams would exceed this limit (except in the case of a process upset or malfunction, in which case the fuel gases would be exempt from meeting the limit), continuous monitoring of these streams is unnecessary. We have divided these streams into four overall categories, as specified in proposed 40 CFR 60.105(a)(4)(iv)(A) through (D). The first category includes pilot gas flames, which are fairly insignificant sources. Although previous determinations effectively excluded these gases from the requirements of the rule, we believe it is good air pollution control practice to fire pilot lights with natural gas or treated fuel gas. However, even when considering the pilot flame as part of the fuel gas combustion device, the potential for sulfur oxide emissions from these sources is insignificant and it is not cost-effective to require continuous monitoring of these gas streams. Therefore, we are changing in the monitoring requirements that monitoring of pilot flame fuel gas is not required.

The second category includes gas streams that meet commercial-grade product specifications with a sulfur content of 30 ppmv or less. Placing a limit on the sulfur content of the products that we are proposing to exempt from monitoring ensures that only low-sulfur products are excluded. The 30 ppmv limit for commercial-grade gas products was selected because it provides a sufficient margin of safety to ensure continuous compliance with the proposed annual average H

2

S concentration limit of 60 ppmv regardless of normal fluctuations in the composition of commercial grade products.

We are requesting comment on the appropriateness of an additional exemption for gas streams that were generated from certain commercial-grade liquid products (e.g., displaced vapors from a storage tank or loading rack for gasoline or diesel fuel). The most straightforward approach would be to exempt gas streams associated with commercial liquid products that contain sulfur below some specified weight percent level. For example, we expect that most of the sulfur-containing compounds in gasoline meeting the Tier 2 sulfur standards or in diesel fuel

meeting the low-sulfur diesel fuel standards have high molecular weights and low vapor pressures such that gas streams in equilibrium with them would have sulfur contents below the proposed 30 ppmv level. To confirm this assumption, we are asking for data on the typical concentrations and vapor pressures of the most prevalent mercaptans, thiophenes, and other sulfur-containing compounds in these or other commercial liquid products.

We would use these data to calculate the corresponding vapor phase concentrations of gas streams in equilibrium with the liquid products using Raoult's Law. Given the extremely low concentrations of the sulfur-containing compounds in the liquid products, we are also seeking comment on whether Raoult's Law gives a realistic estimate of their vapor phase partial pressures. We are also interested in any test data to support this approach, and we are interested in any other approaches to develop an exemption for gas streams associated with commercial-grade liquid products.

The third category includes fuel gases produced by process units that are intolerant of sulfur contamination. There are a few process units within a refinery whose operation is dependent on keeping the sulfur content low. If there is too much sulfur in the gas streams entering these units, the process units could malfunction. Specifically, the methane reforming unit in the hydrogen plant, the catalytic reforming unit, and the isomerization unit are intolerant of sulfur in the process streams; therefore, these streams are treated to remove sulfur prior to processing in these units. Fuel gases subsequently formed in these process units are low in sulfur because the process feedstocks are necessarily low in sulfur. As such, we find that requiring continuous monitoring of the H

2

S content in these gas streams or requiring each individual refinery to develop and implement an alternative monitoring plan (AMP) is unnecessary and creates needless obstacles to using the produced fuel gas directly in the heaters associated with these process units. We are asking for comment on whether fuel gas is generated from any other process units that are intolerant of sulfur. Comments recommending the exemption of fuel gas streams from other units should identify the problems sulfur cause in the unit, procedures used to reduce sulfur in the gas stream before it is processed in the unit, and the expected sulfur content of the outlet fuel gas stream.

For all of the above low-sulfur streams that an owner or operator determines are exempt from all monitoring requirements, the owner or operator must document which of the exemptions applies to each stream. If the refinery operations associated with an exempt stream change, the owner or operator must document the change and determine whether the stream continues to be exempt. If the refinery operations or the composition of an exempt stream change in such a way that the stream is no longer exempt from monitoring, the owner or operator must begin continuous monitoring within 15 days after the change occurs.

In addition, we are proposing a standardized, streamlined procedure to exempt from continuous monitoring streams that an owner or operator can demonstrate are inherently low-sulfur (i.e., consistently 5 ppmv or less H

2

S) following the procedures specified in proposed 40 CFR 60.105(b). The information that an owner or operator must provide to EPA is similar to the information and items needed to apply for an AMP, as described in the EPA document “Alternative Monitoring Plan for NSPS Subpart J Refinery Fuel Gas.” In general, once an AMP is approved for an affected source, the owner or operator must continue to monitor the stream, although a methodology other than a continuous monitor may be used. For this specific exemption, however, once an application to demonstrate that a stream is inherently low-sulfur is approved by EPA, that stream is exempt from monitoring until there is a change in the refinery operation that affects the stream or the stream composition changes. If the sulfur content of the stream changes but is still within the range of concentrations included in the original application, the owner or operator will conduct H

2

S testing on a grab sample as proof and record the results of the test. If the sulfur content of the stream changes such that the sulfur concentration is outside the range provided in the original application, the owner or operator must submit a new application that must be approved in order for the stream to continue to be exempt from continuous monitoring. If a new application is not submitted, the owner or operator must begin continuous monitoring within 15 days.

B. How is EPA proposing to amend definitions?

We are proposing to amend the definition of “Claus sulfur recovery plant” in 40 CFR 60.101(i). These changes would clarify that the SRP may consist of multiple units, and the types of units that are part of a SRP would be listed within the definition. Note that sulfur pits would be included as one of the units, which is consistent with the Agency's current interpretation of the existing definition.

In conjunction with this amendment, we are also proposing to amend the definitions of “oxidation control system” and “reduction control system” in 40 CFR 60.101(j) and 40 CFR 60.101(k), respectively. The amended definitions would specify that the oxidized or reduced sulfur is recycled to the beginning of a sulfur recovery train within the SRP and are consistent with the proposed definitions in 40 CFR 60.101a of subpart Ja. This clarification would ensure that thermal oxidizers that convert the sulfur to SO

2

but do not recycle and recover the oxidized sulfur are not considered oxidation control systems.

C. How is EPA proposing to revise the coke burn-off equation?

The current equation for calculating coke burn-off rate in 40 CFR 60.106(b)(3) assumes that each fluid catalytic cracking unit is using air with 21 percent O

2

. However, there are some fluid catalytic cracking units that use O

2

-enriched air, and for these units, the current equation is not completely accurate. Equation 1 in 40 CFR 63.1564(b)(4)(i) of the National Emission Standards for Hazardous Air Pollutants for Petroleum Refineries: Catalytic Cracking Units, Catalytic Reforming Units, and Sulfur Recovery Units (40 CFR part 63, subpart UUU) includes an additional term to account for the use of an O

2

-enriched air stream. For accuracy in the calculation of the coke burn-off rate, we are proposing to revise the coke burn-off rate equation in 40 CFR 60.106(b)(3) to be consistent with the equation in 40 CFR 63.1564(b)(4)(i). This revision also includes changing the constant values and the units of the resulting coke burn-off rate from Megagrams per hour (Mg/hr) and tons per hour (tons/hr) to kilograms per hour (kg/hr) and pounds per hour (lb/hr).

D. What miscellaneous corrections are being proposed?

See Table 1 of this preamble for the miscellaneous technical corrections not previously described in this preamble that we are proposing throughout 40 CFR part 60, subpart J.

Table 1.—Proposed Technical Corrections to 40 CFR Part 60, Subpart J

Section

Proposed technical correction and reason

60.100

Replace instances of “construction or modification” with “construction, reconstruction, or modification.”

60.100(b)

Replace “except Claus plants of 20 long tons per day (LTD) or less” with “except Claus plants with a design capacity of 20 long tons per day (LTD) or less” to clarify that the size cutoff is based upon design capacity and sulfur content in the inlet stream rather than the amount of sulfur produced.

60.100(b)

Insert ending date for applicability of 40 CFR part 60, subpart J; sources beginning construction, reconstruction, or modification after this date will be subject to 40 CFR part 60, subpart Ja.

60.101

Rearrange definitions alphabetically for ease in locating a specific definition.

60.102(b)

Replace “g/MJ” with “grams per Gigajoule (g/GJ)” to correct units.

60.104(b)(1)

Replace “50 ppm by volume (vppm)” with “50 ppm by volume (ppmv)” for consistency in unit definition.

60.104(b)(2)

Add “to reduce SO

2

emissions” to the end of the phrase “Without the use of an add-on control device” at the beginning of the paragraph to clarify the type of control device to which this paragraph refers.

60.105(a)(3)

Add “either” before “an instrument for continuously monitoring” and replace “except where an H

2

S monitor is installed under paragraph (a)(4)” with “or monitoring as provided in paragraph (a)(4)” to more accurately refer to the requirements of § 60.105(a)(4) and clarify that there is a choice of monitoring requirements.

60.105(a)(3)(iv)

Replace “accurately represents the SO

2

emissions” with “accurately represents the SO

2

emissions” to correct a typographical error.

60.105(a)(4)

Replace “In place” with “Instead” at the beginning of this paragraph to clarify that there is a choice of monitoring requirements.

60.105(a)(8)

Replace “seeks to comply with § 60.104(b)(1)” with “seeks to comply specifically with the 90 percent reduction option under § 60.104(b)(1)” to clearly identify the emission limit option to which the monitoring requirement in this paragraph refers.

60.105(a)(8)(i)

Change “shall be set 125 percent” to “shall be set at 125 percent” to correct a grammatical error.

60.106(e)(2)

Replace the incorrect reference to 40 CFR 60.105(a)(1) with a correct reference to 40 CFR 60.104(a)(1).

60.107(c)(1)(i)

Replace both occurrences of “50 vppm” with “50 ppmv” for consistency in unit definition.

60.107(f)

Redesignate current 40 CFR 60.107(e) as 40 CFR 60.107(f) to allow space for a new paragraph (e).

60.107(g)

Redesignate current 40 CFR 60.107(f) as 40 CFR 60.107(g) to allow space for a new paragraph (e).

60.108(e)

Replace the incorrect reference to 40 CFR 60.107(e) with a correct reference to 40 CFR 60.107(f).

60.109(b)(2)

Add a reference to 40 CFR 60.106(e)(3) to specify that determining whether a fuel gas stream is low-sulfur may not be delegated to States.

60.109(b)(3)

Redesignate current 40 CFR 60.109(b)(2) as 40 CFR 60.109(b)(3) to allow space for a new paragraph (b)(2).

V. Rationale for the Proposed Standards (40 CFR part 60, subpart Ja)

A. What is the performance of control technologies for fluid catalytic cracking units?

1. PM Control Technologies

Filterable PM emissions from fluid catalytic cracking units are predominately fine catalyst particles generated from the mechanical grinding of catalyst particles as the catalyst is continuously recirculated between the fluid catalytic cracking unit and the catalyst regenerator. Control of PM emissions from fluid catalytic cracking units relies on the use of post-combustion controls to remove solid particles from the flue gases. Electrostatic precipitators (ESP) and wet scrubbers are the predominant technologies used to control PM from fluid catalytic cracking units. Either of these PM control technologies can be designed to achieve overall PM collection efficiencies in excess of 95 percent.

Electrostatic Precipitator (ESP).

An ESP operates by imparting an electrical charge to incoming particles, and then attracting the particles to oppositely charged metal plates for collection. Periodically, the particles collected on the plates are dislodged in sheets or agglomerates (by rapping the plates) and fall into a collection hopper. The normal PM control efficiency range for an ESP is between 90 and 99+ percent. One of the major advantages of an ESP is that it operates with essentially little pressure drop in the gas stream. They are also capable of handling high temperature conditions.

Wet Scrubbers.

Wet scrubbers use a water spray to coat and agglomerate particles entrained in the flue gas. To improve wetting of fine particulates, either enhanced spray nozzles or venturi acceleration is used. The wetted particles are then removed from the flue gas through centrifugal separation. Wet scrubbers have similar collection efficiencies as dry ESP (90 to 98 percent), but they are also effective in removing SO

2

emissions. Wet scrubbers may also be more effective in controlling condensable PM as they often use water quench and thereby operate at lower temperatures than ESP used to control fluid catalytic cracking units. Wet scrubbers are generally more costly to operate than ESP due to higher pressure drops across the control device and because of water treatment and disposal costs. However, they become economically viable if significant SO

2

emissions reductions are also needed.

Fabric Filters.

A fabric filter collects PM in the flue gases by passing the gases through a porous fabric material. The buildup of solid particles on the fabric surface forms a thin, porous layer of solids, which further acts as a filtration medium. Gases pass through this cake/fabric filter, and all but the finest-sized particles are trapped on the cake surface. Collection efficiencies of fabric filters can be as high as 99.99 percent. Fabric filters tend to be more efficient for fine particles (those less than 2.5 microns in diameter) than ESP or wet scrubbers.

The primary concern with fabric filters are maintenance requirements of the baghouses given the long run times of typical fluid catalytic cracking units. Small process upsets (e.g., pressure changes) in the fluid catalytic cracking unit and regenerator system can send high concentrations of particles to the control system. These particles would likely blind the filter bags, causing a shut-down of the unit to replace the filter bags. Wet scrubbers and ESP can more easily accommodate and control high concentrations of particles.

2. SO

2

Control Technologies

During combustion, sulfur compounds present in the deposited coke are predominately oxidized to gaseous SO

2

. One approach to controlling SO

2

emissions from catalytic cracking units is to limit the maximum sulfur content in the feedstock to the

catalytic cracking unit. This can be accomplished by processing crude oil that naturally contains low amounts of sulfur or a feedstock that has been pre-treated to remove sulfur (i.e., hydrotreatment or hydrodesulfurization). A second approach is to use a post-combustion control technology that removes SO

2

from the flue gases. These technologies rely on either absorption or adsorption processes that react SO

2

with lime, limestone, or another alkaline material to form an aqueous or solid sulfur by-product. A third approach is the use of catalyst additives, which capture sulfur oxides in the regenerator and return them to the fluid catalytic cracking reactor where they are transformed to H

2

S that is ultimately exhausted to the SRP.

Feedstock Selection or Pre-Treatment.

The SO

2

emissions from the fluid catalytic cracking unit are directly related to the amount of sulfur deposited on the catalyst particles in the riser and reactor section of the unit. The amount of sulfur deposited on the catalyst is a function of both the amount of sulfur in the feedstocks and the relative composition of the sulfur-containing compounds in the feedstocks (mercaptans, thiosulfates). As the concentration of sulfur in the feedstocks is reduced, the SO

2

emissions from the regenerator portion of the unit are also reduced. Therefore, if a refinery processes “sweet” crude (oil naturally low in sulfur) or if a refinery removes sulfur from the feedstocks of the fluid catalytic cracking unit, the SO

2

emissions from the catalyst regenerator will be lower than from refineries that process feedstocks that have higher sulfur content. At a petroleum refinery, the primary means of removing sulfur compounds in the liquid feedstocks is catalytic hydrotreatment. Hydrotreatment typically reduces the sulfur content in process streams to between 20 and 1,000 parts per million by weight.

Alkali Wet Scrubbing.

The SO

2

in a flue gas can be removed by reacting the sulfur compounds with a solution of water and an alkaline chemical to form insoluble salts that are removed in the scrubber effluent. Wet scrubbing processes used to control SO

2

are generally termed flue-gas desulfurization (FGD) processes. The normal SO

2

control efficiency range for SO

2

scrubbers is 80 percent to 90 percent for low efficiency scrubbers and 90 percent to 99 percent for high efficiency scrubbers. In recent fluid catalytic cracking unit applications, control guarantees of 25 ppmv SO

2

are commonly provided by FGD suppliers.

Spray Dryer Adsorption.

An alternative to using wet scrubbers is to use spray dryer adsorber (SDA) technology. A SDA operates by the same principle as alkali wet scrubbing, except that instead of a bulk liquid (as in wet scrubbing) the flue gas containing SO

2

is contacted with fine spray droplets of hydrated lime slurry in a spray dryer vessel. This vessel is located downstream of the air heater outlet where the gas temperatures are in the range of 120 °C to 180 °C (250 °F to 350 °F). The SO

2

is absorbed in the slurry and reacts with the hydrated lime reagent to form solid calcium sulfite and calcium sulfate. The water is evaporated by the hot flue gases and forms dry, solid particles containing the reacted sulfur. Most of the SO

2

removal occurs in the spray dryer vessel itself, although some additional SO

2

capture has also been observed in downstream particulate collection devices. The SO

2

removal efficiencies of new lime spray dryer systems are generally greater than 90 percent. Only one refinery has ever used an SDA to control SO

2

from its fluid catalytic cracking unit; this system has since been removed in favor of feedstock hydrotreatment.

Catalyst Additives.

One common method used by refineries to reduce SO

2

emissions from the fluid catalytic cracking unit is the use of catalyst additives (typically various types of metal oxides). The metal oxide reacts with some of the SO

3

in the catalyst regenerator to form a metal sulfate. The metal sulfate is then returned to the cracking unit where the sulfur is converted to a metal sulfide and then to H

2

S and the original metal oxide. The H

2

S is subsequently recovered in the SRP, and the metal oxide returns to the catalyst regenerator to repeat the process. The control efficiency of catalyst additives is difficult to assess, but is generally around 50 percent (ranging from 20 to 70 percent, depending on the application).

3. NO

X

Control Technologies

NO

X

are formed in a catalyst regenerator (and downstream CO boiler, if present) by the oxidation of molecular nitrogen (N

2

) in the combustion air and any nitrogen compounds contained in the fuel (

i.e.

, thermal NO

X

and fuel NO

X

). The formation of NO

X

from nitrogen in the combustion air is dependent on two conditions occurring simultaneously in the unit's combustion zone: high temperature and an excess of combustion air. Under these conditions, significant quantities of NO

X

are formed, regardless of the fuel type burned. There are several NO

X

emission control strategies that can be considered combustion controls (e.g., low NO

X

burners or flue gas recirculation) that reduce the amounts of NO

X

formed during combustion. These control technologies are primarily applicable to incomplete combustion fluid catalytic cracking units controlled by CO boilers. As there is limited or no direct flame in the catalyst regenerator during normal operations, these control strategies may be limited for complete combustion fluid catalytic cracking units. Most post-combustion control technologies involve converting the NO

X

in the flue gas to N

2

and water using either a process that requires a catalyst (called selective catalytic reduction (SCR)) or a process that does not use a catalyst (called selective noncatalytic reduction (SNCR)). A recently developed post-combustion technology (LoTOx

TM

) uses ozone to oxidize NO

X

to nitric pentoxide, which is water soluble and easily removed in a water scrubber.

NO

X

Combustion Controls.

Flue gas recirculation (FGR) uses flue gas as an inert material to reduce flame temperatures. In a typical FGR system, flue gas is collected from the heater or stack and returned to the burner via a duct and blower. The addition of flue gas with the combustion air reduces the O

2

content of the inlet air stream to the burner. The lower O

2

level in the combustion zone reduces flame temperatures which in turn reduces NO

X

emissions. The normal NO

X

control efficiency range for FGR is 30 percent to 50 percent. When coupled with low-NO

X

burners (LNB), the control efficiency increases to 50-72 percent.

LNB technology utilizes advanced burner design to reduce NO

X

formation through the restriction of O

2

, flame temperature, and/or residence time. The two general types of LNB are staged fuel and staged air burners. Staged fuel LNB are particularly well suited for boilers and process heaters burning process and natural gas which generate higher thermal NO

X

. The estimated NO

X

control efficiency for LNB when applied to petroleum refining fuel burning equipment is generally around 40 percent.

One NO

X

combustion control technique that is applicable to complete combustion fluid catalytic cracking units is the use of catalyst additives and/or combustion promoters. The control efficiency of these additives varies from 10 to 50 percent.

Selective Catalytic Reduction (SCR) Technology.

The SCR process uses a catalyst with ammonia (NH

3

) to reduce the nitrogen oxide (NO) and nitrogen dioxide (NO

2

) in the flue gas to N

2

and water. Ammonia is diluted with air or

steam, and this mixture is injected into the flue gas upstream of a metal catalyst bed that typically is composed of vanadium, titanium, platinum, or zeolite. The SCR catalyst bed reactor is usually located between the economizer outlet and air heater inlet where temperatures range from 230 °C to 400 °C (450 °F to 750 °F). The SCR technology is capable of NO

X

reduction efficiencies of 90 percent or higher.

Selective Noncatalytic Reduction (SNCR) Technology.

An SNCR process is based on the same basic chemistry of reducing the NO and NO

2

in the flue gas to N

2

and water, but it does not require the use of a catalyst to promote these reactions. Instead, the reducing agent is injected into the flue gas stream at a point where the flue gas temperature is within a specific temperature range of 870°C to 1,090°C (1,600°F to 2,000°F). The NO

X

reduction levels for SNCR are in the range of approximately 30 to 50 percent.

LoTOx

TM

Technology

. The LoTOx

TM

process (i.e., low-temperature oxidation) is a patented technology that uses ozone to oxidize NO

X

to nitric pentoxide and other higher order NO

X

, all of which are water soluble and easily removed from exhaust gas in a wet scrubber. The system operates optimally at temperatures below 300°F. Thus, ozone is injected after scrubber inlet quench nozzles and before the first level of scrubbing nozzles. Outlet NO

X

emission levels have been reduced to less than 20 ppmv, and often as low as 10 ppmv, when inlet NO

X

concentrations ranged from 50 to 200 ppmv.

B. What is the performance of control technologies for fuel gas combustion?

Refinery fuel gas is generally used in process heaters and boilers to meet the energy demands of the refinery. Excess refinery fuel gas is typically combusted using flares. Flares also serve an important safety function to destroy organics and convert H

2

S to SO

2

during process upsets and malfunctions.

Over the past several years, many refineries have reduced flaring episodes by adding flare gas recovery systems and/or by changing their start-up and shutdown procedures to limit flaring. Installing a flare gas recovery system and implementing new start-up and shutdown procedures are expected to reduce VOC, sulfur oxides, and NO

X

emissions from flares. Improved amine scrubbing systems are expected to reduce sulfur oxide emissions from all fuel gas combustion systems. In addition, excess capacity in the SRP will help to minimize sour gas flaring that might be caused by a malfunction in the SRP. Each of these “control” techniques are described in the following paragraphs.

Flare Gas Recovery Systems.

Flare gas recovery systems recover fuel gas from the flare gas header prior to the flare's liquid seal. A flare gas recovery system consists of a compressor, separator, and process controls (to maintain slight positive pressure on the flare header). Flare gas recovery systems are typically designed to recover fuel gas from miscellaneous processes that might regularly be relieved to the flare header system and can effectively recover 100 percent of these fuel gases. However, flare gas recovery systems cannot recover large quantities of fuel gas that might be suddenly released to the flare header system as a result of a process upset or malfunction. These gases would still be flared as necessary to maintain the integrity of the process units and the safety of the plant personnel.

Modified Start-up and Shutdown Procedures.

Although flaring is necessary to ensure safety during process upsets and malfunctions, start-up and shutdown procedures can be designed so as to minimize flaring. For example, depressurization of process vessels can be performed more slowly so as to not overwhelm the fuel gas needs of the refinery and/or the capacity of the flare gas recovery system. Depending on the number of units being shut down at a given time, nearly 100 percent of flaring can be eliminated during start-up and shutdown. There are cases, such as emergency shutdowns for safety reasons or approaching hurricanes, where the timing of the shutdown and the magnitude of the number of processes needing to be shut down would warrant the use of flaring. However, modified procedures should be able to eliminate flaring associated with process start-ups and shutdowns due to routine maintenance of select processes.

Amine Scrubbers.

Amine scrubber systems remove H

2

S and other impurities from sour gas. Lean amine solution absorbs the H

2

S from the sour gas in an absorption tower. The acid gas is removed from the rich amine solution in a stripper, or still column. The resulting lean amine is recirculated to the absorption tower, and the stripped H

2

S is generally sent to the SRP. Vendors generally provide redundant pumps to ensure continuous operation of the system. Some refineries choose to store a day's worth of lean amine solution in case the stripper fails; this allows the continuous operation of the absorption tower. This option also requires adequate empty storage space for the rich amine solution produced by the absorption tower while the stripper is out of service.

Redundant Sulfur Recovery Capacity.

When a sulfur recovery unit (SRU) malfunctions, the sour gas is typically flared to convert the highly toxic H

2

S to less toxic SO

2

. As many SRU recover more than 20 long tons of elemental sulfur per day, even short sulfur recovery process upsets can result in several tons of SO

2

emissions. Furthermore, refineries often operate multiple Claus sulfur recovery processes in parallel. Having an extra Claus sulfur recovery train can dramatically reduce the likelihood of sour gas flaring. Depending on the severity of the process upset, having a redundant SRU can reduce these large SO

2

releases by as much as 100 percent.

C. What is the performance of control technologies for process heaters?

The mechanisms by which NO

X

are formed in process heaters are the same as for their formation in catalyst regenerators. The possible control options are also the same. See section V.A.3 of this preamble for a discussion of these formation mechanisms and control technologies.

D. What is the performance of control technologies for sulfur recovery systems?

Sulfur recovery (the conversion of H

2

S to elemental sulfur) is typically accomplished using the modified-Claus process. In the Claus unit, one-third of the H

2

S is burned with air in a reaction furnace to yield SO

2

. The SO

2

then reacts reversibly with H

2

S in the presence of a catalyst to produce elemental sulfur, water, and heat. This is a multi-stage catalytic reaction in which elemental sulfur is removed between each stage, thereby driving the reversible reaction towards completion. The gas from the final condenser of the Claus unit (referred to as the “tail gas”) consists primarily of inert gases with less than 2 percent sulfur compounds. Additionally, the sulfur recovery pits used to store the recovered elemental sulfur also have a potential for fugitive H

2

S emissions. Typically a Claus unit recovers approximately 94 to 97 percent of the inlet sulfur load as elemental sulfur.

There are some methods that extend the Claus reaction to improve the overall sulfur collection efficiency of the SRP. For example, the Superclaus® SRU is similar to the Claus unit. It contains a thermal stage, followed by three to four catalytic reaction stages. The first two or three catalytic reactors use the Claus catalyst, while the last reactor uses a selective oxidation catalyst. The

catalyst in the last reactor oxidizes the H

2

S to sulfur at a very high efficiency, recovering 99 percent of the incoming sulfur.

There are a few refineries that operate non-Claus type SRU. All of the refineries that use non-Claus SRU technologies have very low sulfur production rates (2 LTD or less). There are several different trade names for these “other” types of SRU, such as the LoCat®, Sulferox®, and NaSH processes. These processes can achieve sulfur recovery efficiencies of 99 percent or more, although they typically yield a sulfur product that has limited market value because the sulfur content is much lower than in the sulfur product from a Claus unit (50 to 70 percent sulfur compared to 99.9 percent sulfur from the Claus process).

The primary means of reducing sulfur oxide emissions from the SRU is to employ a tail gas treatment unit that recovers the sulfur compounds and recycles them back to the inlet of the Claus treatment train. There are three basic types of tail gas treatment units: (1) Direct amine adsorption of the Claus tail gas; (2) catalytic reduction of the tail gas to convert as much of the tail gas sulfur compounds to H

2

S (coupled with amine adsorption or Stretford solution eduction); and (3) oxidative tail gas treatment systems to convert the Claus tail gas sulfur compounds to SO

2

(coupled with an SO

2

recovery system).

Direct Amine Adsorption.

Direct amine adsorption of the Claus tail gas is the least efficient of the tail gas treatment methods because only about two-thirds of the sulfur in the direct Claus tail gas is amenable to scrubbing (i.e., in the form of H

2

S). Direct amine adsorption is therefore expected to increase the overall sulfur recovery efficiency of the sulfur plant to approximately 99 percent. However, direct amine adsorption alone is generally not expected to reduce sulfur oxide concentrations to below 250 ppmv (i.e., enough to meet the existing NSPS emission limits for Claus units greater than 20 LTD).

Reductive Tail Gas Catalytic Systems.

The most common reductive tail gas catalytic systems in use at refineries include: (1) The Shell® Claus Offgas Treatment (SCOT) unit; (2) the Beavon/amine system; and (3) the Beavon/Stretford system. Each of these systems consist of a catalytic reactor to convert the sulfur compounds remaining in the Claus tail gas to H

2

S and an H

2

S recovery system (an amine scrubber or a Stretford solution) to strip the H

2

S from the tail gas. The recovered H

2

S is then recycled to the front of the Claus unit. The overhead of the amine scrubber or Stretford unit (caustic scrubber) may be vented to the atmosphere or incinerated to convert any remaining H

2

S or other reduced sulfur compounds to SO

2

. The total sulfur recovery efficiency of a Claus/catalytic tail gas treatment train is expected to be 99.7 to 99.9 percent.

Oxidative Tail Gas Treatment Systems.

The Wellman-Lord is the only oxidative tail gas treatment system used in the United States. The Wellman-Lord process uses thermal oxidation followed by scrubbing with a sodium sulfite and sodium bisulfite solution to remove SO

2

. The rich bisulfite solution is sent to an evaporator-recrystallizer where the bisulfite decomposes to SO

2

and water and sodium sulfite is precipitated. The recovered SO

2

is then recycled back to the Claus plant for sulfur recovery. The total sulfur recovery efficiency of a Claus/oxidative tail gas treatment train is expected to be 99.7 to 99.9 percent.

E. How did EPA determine the proposed standards for new petroleum refining process units?

Four sources of information were considered in reviewing the appropriateness of the current NSPS requirements for new sources: (1) Source test data from recently installed control systems; (2) applicable State and local regulations; (3) control vendor emission control guarantees; and (4) consent decrees. (A significant number of refineries, representing about 77 percent of the national refining capacity, are subject to consent decrees that limit the emissions from subpart J process units.) Once we identified potential emission limits for various process units, we evaluated each limit in conjunction with control technology, costs, and emission reductions to determine BDT for each process unit.

The cost methodology incorporates the calculation of annualized costs and emission reductions associated with each of the options presented. Cost-effectiveness is the annualized cost of control divided by the annual emission reductions achieved. Incremental cost-effectiveness refers to the difference in annualized cost from one option to the next divided by the difference in emission reductions from one option to the next. For NSPS regulations, the standard metric for expressing costs and emission reductions is the impact on all affected facilities accumulated over the first 5 years of the regulation. Details of the calculations can be found in the public docket. Our BDT determinations took all relevant factors into account, including cost considerations which were generally consistent with other Agency decisions.

1. Fluid Catalytic Cracking Units

Particulate Matter (PM) and Sulfur Dioxide (SO

2

).

In order to determine the appropriate emission limits for PM and SO

2

, we evaluated PM and SO

2

limits in conjunction with one another. One of the reasons for this is that wet scrubbers control both PM and SO

2

emissions, and refineries will decide whether to choose a wet scrubber as opposed to an ESP with catalyst additives based on both the PM and the SO

2

emission limit to be met.

Currently, 40 CFR part 60, subpart J limits PM emissions from the fluid catalytic cracking unit to 1.0 kg/Mg of coke burn-off. The limit applies to filterable PM as measured by Method 5B or 5F in 40 CFR part 60, Appendix A. It excludes condensable PM such as sulfuric acid (under Method 5B), sulfates that condense at temperatures greater than 320 °F (under Method 5F), and all other condensables (using either Method). The measurement of condensable PM is important to EPA's goal of reducing ambient air concentrations of fine PM. Since promulgation of Method 202 in 1991, EPA has been working to overcome problems associated with the accuracy of Method 202 and will promulgate improvements to the method in the future. The existing NSPS also requires opacity, as measured using a continuous opacity monitoring system, to be no more than 30 percent.

The current standards in 40 CFR part 60, subpart J for SO

2

include three alternative formats: (1) If using an add-on control device, reduce SO

2

emissions by at least 90 percent or to less than 50 ppmv, (2) if not using an add-on control device, limit sulfur oxides emissions (calculated as SO

2

) to no more than 9.8 kg/Mg of coke burn-off, or (3) process in the fluid catalytic cracking unit fresh feed that has a total sulfur content no greater than 0.30 percent by weight. The 90 percent reduction, 9.8 kg/Mg, and 0.3 percent feed sulfur formats were determined to be equivalent for a unit operating with a feed that contains 3.5 percent sulfur by weight before implementing a control measure.

In reviewing the PM and SO

2

emission limits, we evaluated five combined options and a baseline. The baseline is considered to be the current requirements, as described in the two previous paragraphs. The first option is to maintain the existing subpart J standard for PM and provide only the 50 ppmv concentration limit for SO

2

. The additional options are a range of emission limits coupled with a change in the compliance test method to Method 5 to measure a portion of the

condensable PM. The second option is to combine Method 5 with the existing 1.0 kg/Mg coke burn-off performance level, and a third option is to lower the PM emission limit to 0.5 kg/Mg. Both the second and third options include an SO

2

limit of 50 ppmv. A fourth option includes the PM limit of 0.5 kg/Mg presented in the third option and a lower SO

2

limit of 25 ppmv. The fifth option is to lower the PM emission limit to 0.15 kg/Mg with an SO

2

limit of 25 ppmv. Costs and emission reductions for each option were estimated as the increment between complying with subpart J and subpart Ja.

Option 1 includes the same emissions and requirements for PM as the current 40 CFR part 60, subpart J. For SO

2

, this option excludes the alternative compliance options of meeting a higher emission limit without an SO

2

control device or meeting a limit on the sulfur content of the fresh feed. These two alternatives are less stringent than the outlet concentration limit, and available information indicates the concentration limits are achievable. An advantage of the proposed concentration limit is that ongoing compliance can be directly measured using a CEMS. The impacts of this option are limited to the impacts of removing those alternative compliance options for SO

2

and are presented in Table 2 to this preamble. To comply with Option 1 (i.e., meet the 50 ppmv limit for SO

2

) we expect that the fraction of new sources choosing wet scrubbers instead of ESP would be greater than under the existing subpart J. Filterable PM emissions are assumed to be the same for both types of control devices because the PM performance levels are the same under both option 1 and the baseline subpart J requirements. However, because condensable PM emissions are lower from wet scrubbers than from ESP, this shift in the ratio of wet scrubbers to ESP would also result in an estimated reduction in total PM emissions of 17 tons per year, as shown in Table 2 to this preamble.

Option 2 includes the same emission limit as current subpart J for PM but requires compliance using Method 5 rather than Method 5B or Method 5F. As noted above, Methods 5B and 5F exclude all PM that condenses at temperatures below 320°F, and Method 5F also excludes sulfates that condense at temperatures greater than 320°F. The PM measured by Method 5 includes filterable PM that condenses above 250°F in the front half of the Method 5 sampling train. Thus, the estimated PM emission reductions achieved by this option equal the amount of sulfates and other condensable PM between 250°F and 320°F that would be measured by Method 5 but not Method 5B or 5F. The baseline emissions were estimated assuming Method 5B is used for wet scrubbers and Method 5F is used for ESP. For SO

2

, Option 2 includes the same emission limit as described in Option 1, and the estimated SO

2

emission reductions are also the same. The impacts of this option are presented in Table 2 to this preamble.

Option 3 lowers the PM limit to 0.5 kg/Mg coke burn, again using Method 5, and includes the same emission limit as described in Option 1 for SO

2

. The existing NSPS limit was based on control with ESP. Those ESP were rated at efficiencies of only 85 to 90 percent. More recently installed ESP have greater specific plate area, which should result in better control efficiencies. In addition, many refineries have installed wet scrubbers to control both PM and SO

2

. At petroleum refineries, wet scrubbers typically perform as well as, if not better than, ESP. Available test data indicate that at least one ESP and one wet scrubber are reducing total filterable PM to 0.5 kg/Mg of coke burn or less, as measured by Method 5-equivalent test methods. Based on this information, both ESP and wet scrubbers can achieve PM emission levels below the level of the existing PM standard, and a lower standard for new units is technically feasible. The impacts of this option are presented in Table 2 to this preamble.

Option 4 includes the same PM limit as Option 3, and the discussion presented for Option 3 applies to Option 4 as well. It also includes a long-term limit for SO

2

of 25 ppmv, averaged over 365 days, in addition to the current subpart J limit of 50 ppmv, averaged over 7 days. These limits have been shown to be readily achievable by flue gas desulfurization systems. Many fluid catalytic cracking units are now subject to consent decrees that require control to these levels. Petroleum refiners typically use wet scrubbers to control SO

2

emissions, and test data indicate that outlet concentrations below 25 ppmv are common. At least one wet scrubber manufacturer also provides performance guarantees to meet a 25 ppmv emission limit. The incremental SO

2

reductions for this option relative to Option 3 are achieved by using catalyst additives in the fluid catalytic cracking units that are assumed to be controlled with ESP; fluid catalytic cracking units controlled with wet scrubbers have the same SO

2

emissions as under Option 3 because wet scrubbers under all options are assumed to achieve SO

2

emissions below 25 ppmv. The impacts of this option are presented in Table 2 to this preamble.

The final option, Option 5, includes a lower PM limit, 0.15 kg/Mg of coke burn, measured using Method 5, and the same SO

2

limits as Option 4. This PM limit is equivalent to the limit of 0.005 gr/dscf required by California's South Coast Air Quality Management District (SCAQMD). To meet this PM limit, we expect that a refinery would need an ESP rather than a wet scrubber because we are unaware of any wet scrubber that is meeting this PM limit (and as in Option 4, catalyst additives in the fluid catalytic cracking unit would be needed to meet the SO

2

limit). In addition, the refinery would likely need ammonia injection to improve the performance of the ESP. Based on test data from at least three fluid catalytic cracking units, ammonia injection improves the control of filterable PM in ESP, but it also produces a considerable amount of condensable PM. Therefore, the estimated total PM reduction for this option is much lower (worse) than the reduction that would be achieved under Option 4. The shift to ESP for all new fluid catalytic cracking units under this option also slightly degrades the estimated SO

2

emissions reduction relative to Option 4 because available data indicate that wet scrubbers achieve lower SO

2

emissions than ESP and catalyst additives. In addition to reduced performance relative to Option 4, the capital and annual costs of this option are considerably higher than for Option 4. The reduced performance of this option relative to Option 4 means that incremental cost-effectiveness is not meaningful for this option. The impacts of this option are presented in Table 2 to this preamble.

Table 2.—National Fifth Year Impacts of Options for PM and SO

2

Limits Considered for Fluid Catalytic Cracking Units Subject to 40 CFR Part 60, Subpart Ja

Option

Capital cost

($1,000)

Total annual cost

($1,000/yr)

Emission

reduction (tons PM/yr)

a

Emission

reduction (tons SO

2

/yr)

Cost-effectiveness ($/ton)

Overall

Incremental

1

500

3,100

17

6,800

460

2

670

3,600

350

6,800

500

1,400

3

40,000

9,200

1,200

7,200

1,100

4,400

4

40,000

9,500

1,200

8,300

1,000

220

5

140,000

30,000

460

7,900

3,600

N/A

a

Both filterable and condensable PM.

Based on our review of performance data and potential impacts, we have determined that control of PM emissions (as measured by Method 5) to 0.5 kg/Mg of coke burn or less and control of SO

2

emissions to 25 ppmv or less averaged over 365 days and 50 ppmv or less averaged over 7 days is BDT for new, reconstructed, or modified fluid catalytic cracking units. The more stringent filterable PM control level in Option 5 is technically achievable, but we rejected this option because it results in higher total PM and SO

2

emissions than Option 4. Option 4 was selected as BDT because it achieves the best performance of the remaining options, and both overall and incremental costs are reasonable.

Table 3 to this preamble shows the impacts of Option 4 for modified and reconstructed sources. Although the impacts of Option 4 are reasonable, we are aware that there is some concern about the ability to retrofit reconstructed and modified sources to meet these emission limits. Specifically, there may be issues with physical space availability, process unit or control device configurations, or other factors that are not adequately included in our impacts analyses. Therefore, we are co-proposing requiring reconstructed and modified units to meet the current standards in 40 CFR part 60, subpart J. We are requesting comment on specific examples, supported by data, of situations that would support this proposed option.

Table 3.—National Fifth Year Impacts of Proposed Option for PM and SO

2

Limits for Reconstructed and Modified Sources

Capital cost ($1,000)

Total annual cost

($1,000/yr)

Emission

reduction

(tons PM/yr)

Emission

reduction

(tons SO

2

/yr)

Cost-effectiveness ($/ton)

31,000

6,200

700

3,700

1,400

Finally, available test data indicate that the two control devices (an ESP and a wet scrubber) that reduce filterable PM to less than 0.5 kg/Mg coke burn (as well as at least one other ESP) also can meet a total PM limit, including condensables, of 1.0 kg/Mg of coke burn (i.e., demonstrate compliance using Method 5 for filterable PM and Method 202 for condensable PM). Condensable sulfates and other condensable compounds measured by Method 5 and Method 202 vary widely, but the average is about 0.5 kg/Mg of coke burn-off. In an attempt to create some incentive to begin measuring condensables using improved Method 202, we are considering establishing an alternative PM limit of 1 kg/Mg coke burn, including condensables. Therefore, we are asking for comments with rationale to either support or reject an alternative PM limit that would be based on both filterable PM and condensable PM.

Carbon Monoxide

. The current standards in 40 CFR part 60, subpart J limit CO emissions to 500 ppmv or less. This limit was established for fluid catalytic cracking units that operate in either “partial combustion” catalyst regeneration mode or “complete combustion” catalyst regeneration mode. In partial combustion mode, relatively large amounts of CO are generated in the regenerator. The resulting CO is then combusted in a CO or waste heat boiler. This operation results in nearly complete combustion of the CO, with outlet concentrations on the order of 25 to 50 ppmv being common. In complete combustion mode the CO emissions from the regenerator are much lower, and a downstream CO or waste heat boiler is impractical. However, complete combustion catalyst regeneration was a recent advance at the time the current NSPS was promulgated; test data were limited at that time, and a CO level of 500 ppmv was estimated to be a practical limit for the technology.

After consideration of available information, we are proposing to retain the current CO standard for new fluid catalytic cracking units. Although test data show CO emissions from complete combustion regenerators can be less than 500 ppmv, the lower levels generally are achieved by operating with higher levels of excess air. Unfortunately, this operation is likely to result in higher NO

X

emissions. If a trade-off is necessary, limiting NO

X

emissions is a higher priority than limiting CO emissions because NO

X

is a precursor to fine PM and ground-level ozone, both of which have more significant health impacts than CO. Available data also indicate that formaldehyde emissions tend to increase with the higher oxidation/combustion conditions needed to reduce CO emissions. Therefore, we determined that control to 500 ppmv or less is still BDT for CO emissions, and the proposed standards are based on this emission limit. Accordingly, the proposed limit for 40 CFR part 60, subpart J poses no additional costs over those incurred to comply with the existing NSPS.

NO

X

.

NO

X

emissions are not subject to control under the existing NSPS in 40 CFR part 60, subpart J. However, several petroleum refiners limit NO

X

emissions based on State regulations and consent decrees. The emission limits to which refineries are subject vary from facility to facility. We evaluated three options

as part of the BDT determination: Outlet NO

X

emission levels of 80 ppmv, 40 ppmv, and 20 ppmv, each averaged over 7 days or less. Each of these limits is technically feasible, but the technology needed to meet them depends on the current NO

X

concentrations in the vented gas streams, which are either uncontrolled or controlled to levels required by existing State and local requirements.

The estimated fifth year emission reductions and costs for each of the options are summarized in Table 4. To estimate impacts for Option 1, we assumed that a few units have current NO

X

emissions below 80 ppmv, and many other units can meet this level with combustion controls (e.g., limiting excess O

2

or using non-platinum catalyst combustion promoters in a complete combustion catalyst regenerator, or using flue gas recirculation or low-NO

X

burners in a CO boiler after a partial combustion catalyst regenerator). Other units with higher uncontrolled NO

X

emissions levels will need to install more costly control technology such as LoTOx

TM

or SCR in order to meet the 80 ppmv option. All units will also incur costs for a continuous NO

X

monitor. The costs for Options 2 and 3 are higher than for Option 1 because the ratio of add-on controls to combustion controls would increase in order to meet the lower limits of 40 and 20 ppmv.

Based on the impacts shown in Table 4, we determined that BDT is option 1, a NO

X

emission limit of 80 ppmv. The costs of option 1 are commensurate with the emission reductions while the more stringent options would impose compliance costs that are not warranted for the emissions reductions that would be achieved as shown by the incremental cost effectiveness impacts shown in table 4. In general, we expect that most sources will be able to meet the NO

X

limit through combustion controls. In cases where add-on controls would be necessary, however, there may be retrofit concerns for modified and reconstructed sources. Therefore, we are co-proposing no new standards for NO

X

emissions on modified or reconstructed sources and are requesting comments on the necessity, feasibility and costs of retrofits to meet the 80 ppmv limit for modified and reconstructed sources.

Table 4.—National Fifth Year Impacts of Options for NO

X

Limits Considered for Fluid Catalytic Cracking Units Subject to 40 CFR part 60, subpart Ja

Option

Total capital cost, $

(millions)

Total annual

cost, $/yr

(millions)

Emission

reduction, tons NO

X

/yr

Cost effectiveness ($/ton)

Overall

Incremental

1

28

7.3

3,500

2,100

2

80

20

5,200

4,200

7,600

3

120

30

5,800

5,500

16,000

Available test data for units controlled with SCR indicate that emissions less than 20 ppmv are continuously achievable when averaged over long periods of time such as 365 days. Although we determined that the average costs to meet such a limit are unreasonable, we are requesting comment on whether there may be a subset of units for which costs would be reasonable to meet lower limits such as 20 or 40 ppmv, averaged over 365 days.

Opacity.

The current standards require fluid catalytic cracking units to meet an opacity limit of 30 percent. This limit was included as a means of identifying failure of the PM control device. This objective is achieved much more effectively by monitoring control device operating parameters or by using a PM CEMS. These monitoring options are included in the proposed standards for PM. Therefore, the proposed standards do not include an opacity emissions limit.

2. Fluid Coking Units

The current NSPS includes no requirements for fluid coking units. There are few fluid coking units at refineries in the U.S., but data in the National Emission Inventory database shows the few existing units are significant sources of PM, SO

2

, and NO

X

emissions. Therefore, we evaluated several options as part of a BDT determination for fluid coking units. All of the options we considered are comparable to options that we considered for fluid catalytic cracking units because of similarities in the function, operation, and emissions of the two types of units.

Particulate Matter and Sulfur Dioxide.

To determine BDT for PM and SO

2

emissions we evaluated two options. Because control technology can reduce both pollutants simultaneously, the options also consider both pollutants. Option 1 is a PM limit of 1.0 kg/Mg coke burn and a short-term SO

2

limit of 50 ppmv, averaged over 7 days; and Option 2 is a PM limit of 0.5 kg/Mg coke burn, a short-term SO

2

limit of 50 ppmv, averaged over 7 days, and a long-term SO

2

limit of 25 ppmv, averaged over 365 days. (Because catalyst additives are not a feasible option for reducing SO

2

from a fluid coking unit, we did not consider the fifth option evaluated for fluid catalytic cracking units.)

The Energy Information Administration (EIA) Refinery Capacity Report 2006 lists six fluid coking units; at least two of these coking units are flexi-coking units that use the coking exhaust as a synthetic fuel gas. Therefore, there are at most four fluid coking units in the United States that could potentially become subject to the standard. Although coking capacity is expected to increase, most new units are expected to be delayed coking units. For this analysis, we assumed that one existing fluid coking unit becomes a modified or reconstructed source in the next 5 years. A wet scrubber is the most likely technology that would be used to meet either Option 1 or Option 2. To estimate the impacts, we estimated costs for a basic wet scrubber to meet Option 1 and an enhanced wet scrubber to meet Option 2. The resulting emission reductions and costs for both of the options are shown in Table 5 to this preamble. The costs for both options are reasonable. Therefore, we determined that BDT is Option 2 which requires technology that reduces PM emissions to 0.5 kg/Mg of coke burn and reduces SO

2

emissions to 50 ppmv, averaged over 7 days, and 25 ppmv, averaged over 365 days. We are proposing standards consistent with these levels.

Table 5.—National Fifth Year Impacts of Options for PM and SO

2

Limits Considered for Fluid Coking Units Subject to 40 CFR Part 60, Subpart Ja

Option

Capital cost

($1,000)

Total annual cost

($1,000/yr)

Emission

reduction

(tons

PM/yr)

Emission

reduction

(tons

SO

2

/yr)

Cost-effectiveness

($/ton)

Overall

Incremental

1

14,000

4,700

1,700

21,000

210

2

14,000

4,800

2,000

21,000

210

120

Nitrogen Oxides (NO

X

).

To determine BDT for NO

X

emissions, we evaluated three options: Outlet NO

X

emission levels of 80 ppmv, 40 ppmv, and 20 ppmv, each averaged over 7 days or less. The specific technology that will be needed to meet these levels will depend on the NO

X

concentration in the exhaust gas stream from uncontrolled fluid coking units. As noted in the discussion above for PM and SO

2

options, we estimated that only one fluid coking unit will be modified or reconstructed in the next 5 years, and there will be no new units constructed. Because each unit is likely to have a different uncontrolled NO

X

concentration in its exhaust stream, we developed impacts for a composite model unit based on a weighted distribution of all the various types of controls (low-efficiency combustion controls, higher efficiency combustion controls, and add-on controls such as LoTox

TM

or SCR). As in the analysis for fluid catalytic cracking units, the ratio of add-on controls to combustions controls increases from Option 1 through Option 3. The results of this analysis are shown in Table 6 to this preamble.

Table 6.—National Fifth Year Impacts Options for NO

X

Limits Considered for Fluid Coking Units Subject to 40 CFR Part 60, Subpart Ja

Option

Total capital cost, $

(millions)

Total annual cost, $/yr

(millions)

Emission

reduction,

(tons

NO

X

/yr)

Cost-effectiveness

($/ton)

Overall

Incremental

1

4.5

0.97

760

1,300

2

9.5

2.1

980

2,200

5,300

3

13

2.9

1,000

2,800

12,000

The costs for option 1 are commensurate with the emission reductions, but the incremental impacts for options 2 and 3 are not reasonable, as shown in Table 6. Based on these potential impacts and available performance data, we have determined that BDT is technology needed to meet an outlet NO

X

concentration of 80 ppmv or less, and we are proposing this emission limit as the performance standard for NO

X

emissions from fluid coking units. However, there are uncertainties in this analysis. For example, if the few existing units are not readily amenable to retrofitting NO

X

controls, the cost and emission reduction impacts might no longer be favorable, and we would conclude that no control is BDT. Therefore, we are co-proposing no new standard for NO

X

emissions from fluid coking units.

3. Sulfur Recovery Plants

Emission limits in the existing NSPS (40 CFR part 60, subpart J) apply to Claus SRP with a capacity greater than 20 LTD. The emission limits are consistent with an overall sulfur recovery efficiency of 99.9 percent (

i.e.

, 250 ppmv SO

2

for the Claus unit followed by oxidative tail gas treatment, and 10 ppmv H

2

S and 300 ppmv total reduced sulfur compounds for a Claus unit followed by reductive tail gas treatment). Although small SRP and non-Claus SRP are not subject to the existing NSPS, they are often subject to control. For example, Texas requires sulfur removal efficiencies of 99.8 percent for SRP with capacities greater than 10 LTD and 96 percent to 98.5 percent for SRP with capacities less than or equal to 10 LTD. In addition, a few consent decrees require 95 percent sulfur recovery for Claus SRP with capacities less than 20 LTD.

To determine BDT we evaluated 4 options. The options are based on various sulfur recovery efficiencies for SRP with capacities less than 20 LTD, and all of the options include the same 99.9 percent efficiency as in the current standards for SRP with capacities greater than 20 LTD. Option 1 is based on 99 percent recovery for SRP with capacities between 10 LTD and 20 LTD, and 95 percent recovery for SRP with capacities less than 10 LTD. Option 2 is based on 99 percent recovery for all SRP with capacities less than 20 LTD. Option 3 is based on 99.9 percent recovery for SRP with capacities between 10 LTD and 20 LTD, and 99 percent recovery for SRP with capacities less than 10 LTD. Option 4 is based on 99.9 percent recovery for all SRP, regardless of size or design. All of the options include 99.9 percent recovery for SRP larger than 20 LTD (both Claus and non-Claus units) because we are not aware of a more effective SO

2

control technology. The 95 percent option is equivalent to the efficiency of a two-stage Claus unit without controls. The 99 percent and 99.9 percent recovery levels are achievable for SRP of all sizes by various types of tail gas treatments, as discussed in section V.D of this preamble.

The estimated fifth year emission reductions and costs for each of the options are summarized in Table 7. These values reflect the impacts only for SRP smaller than 20 LTD because we expect that all non-Claus units will be smaller than 20 LTD and because the impacts for larger Claus units would be the same as to comply with the existing standards in subpart J. The costs for Options 1, 2, and 3 are reasonable. We then evaluated the incremental costs and emission reductions between the options. We found that Option 2 is the most stringent option for which incremental costs are reasonable compared to the incremental emission reduction between the options.

Based on the available performance data and cost considerations, we have concluded that tail gas treatments that achieve 99.9 percent control are still

BDT for SRP with capacities greater than 20 LTD, and tail gas treatments that achieve 99 percent recovery are BDT for SRP with capacities less than 20 LTD. Therefore, we are proposing standards for SO

2

and H

2

S emissions from SRP with capacities larger than 20 LTD that are equivalent to the existing standards, and we are proposing standards for SRP with capacities smaller than 20 LTD that would limit emissions of sulfur to less than 1 percent by weight of the sulfur recovered.

Table 7.—National Fifth Year Impacts of Options for SO

2

Limits Considered for Sulfur Recovery Plants Subject to 40 CFR Part 60, Subpart Ja

Option

Total capital cost, $

(millions)

Total annual cost, $/yr

(millions)

Emission

reduction,

(tons

SO

2

/yr)

Cost-effectiveness

($/ton)

Overall

Incremental

1

0.27

0.14

180

780

2

1.1

0.68

550

1,200

1,500

3

1.9

1.0

590

1,700

8,200

4

4.5

2.3

670

3,400

15,000

4. Process Heaters and Other Fuel Gas Combustion Devices Sulfur Dioxide

The current NSPS in 40 CFR part 60, subpart J limits SO

2

emissions from fuel gas combustion devices by specifying that the H

2

S content of fuel gas must be less than or equal to 230 mg/dscm, averaged over 3 hours (equivalent to 160 ppmv averaged over 3 hours). Alternatively, any fuel gas may be combusted, provided the outlet SO

2

emissions are controlled to no more than 20 ppmv (dry basis, 0 percent excess air). When the current NSPS was promulgated, we concluded that amine scrubbing as well as new processes that use other scrubbing media represented BDT for continuous reduction of H

2

S from fuel gas. The 160 ppmv concentration limit was consistent with good operation of such scrubbing processes. In addition, burning such fuel gas will result in an SO

2

concentration in the exhaust gas of about 20 ppmv.

After consideration of current operating practices, we concluded that amine scrubbing units are still the predominant technology for reduction of H

2

S in fuel gas (and SO

2

emissions from subsequent fuel gas combustion). Considering the variability of the fuel gas streams from various refinery processing units, 160 ppmv also is still a realistic short term H

2

S concentration limit. However, one California Air Quality Management District rule sets a 40 ppmv H

2

S limit in fuel gas (averaged over 4 hours), and several refiners have reported that the typical fuel gas H

2

S concentrations (after scrubbing) are in the same range. Additionally, amine scrubbing technology can be designed and is, in fact, being used to achieve much lower (1 to 5 ppmv) H

2

S concentrations in product gas applications. Based on this information, we concluded that additional SO

2

control could be achieved by requiring SO

2

emission limits with both long-term and short-term averaging periods.

We considered three options for increasing SO

2

control of fuel gas combustion units: Outlet SO

2

emission levels of 10 ppmv, 8 ppmv, and 5 ppmv SO

2

, each averaged over 365 days. Each of the options also includes the same 20 ppmv 3-hour SO

2

concentration limit as in the current NSPS. To achieve each of these options, we expect that petroleum refiners will increase their amine recirculation rates to reduce the H

2

S concentration in the fuel gas. We estimate that meeting the options will increase steam consumption for a typical scrubbing unit by about 5, 7, and 10 percent, respectively. No new equipment or other capital expenditures would be necessary. The estimated fifth-year impacts of each of these options are presented in Table 8 to this preamble. Overall costs for all the options are reasonable compared to the emission reduction achieved. We further evaluated the incremental costs and reductions between the 3 options and found that they were reasonable for Options 1 and 2, while the incremental cost for Option 3 is not.

Table 8.—National Fifth Year Impacts of Options for SO

2

Limits Considered for Process Heaters and Other Fuel Gas Combustion Devices Subject to 40 CFR Part 60, Subpart Ja

Option

Capital cost

($1,000)

Total annual cost

($1,000/yr)

Emission

reduction

(tons

SO

2

/yr)

Cost-effectiveness

($/ton)

Overall

Incremental

1

0

2,000

1,000

1,900

2

0

2,900

1,300

2,200

3,500

3

0

4,100

1,600

2,600

4,700

Based on these impacts and consideration of current operating practices, we concluded that BDT is use of technology that reduces the SO

2

emissions from fuel gas combustion units to 8 ppmv or less averaged over 365 days and 20 ppmv or less averaged over 3 hours. Therefore, we are proposing SO

2

standards consistent with this determination. We are also requesting comment on the proposed long-term concentration limit and the length of the averaging period.

Although the proposed emission limits are based primarily on the fuel gas desulfurization technologies (e.g., amine scrubbing), new process heaters, regardless of fuel type, also would be subject to these emission limits. New process heaters can elect to meet these emission limits by using treated fuel gas, low sulfur distillate fuel oils, or flue gas desulfurization or other SO

2

add-on controls. Considering the low sulfur fuel standards and available control technologies, we believe the 20 ppmv 3-hour average SO

2

emission limit and an

8 ppmv 365-day average emission limit represent the performance of BDT regardless of whether the new process heaters use gaseous or liquid fuels.

The current NSPS allows refineries to demonstrate compliance with fuel gas concentration limits for H

2

S as a surrogate for SO

2

emission limits. This approach is reasonable when H

2

S is the only sulfur-containing compound in the fuel gas because the H

2

S concentration in the fuel gas that is equivalent to the SO

2

concentration in the exhaust from the fuel gas combustion unit can be easily estimated. However, based on available data, we understand that a significant portion of the sulfur in fuel gas from coking units is in the form of methyl mercaptan and other reduced sulfur compounds. These compounds will also be converted to SO

2

in the fuel gas combustion unit, which means the SO

2

emissions will be higher than the amount predicted when H

2

S is the only sulfur-containing compound in the fuel gas. Therefore, for process heaters and other fuel gas combustion devices that burn only fuel gas, we are proposing two alternatives to the SO

2

emission limit. The first option would require measurement of H

2

S if none of the fuel gas is from a coking unit. The H

2

S concentration limits that would be equivalent to the SO

2

emission limits are 160 ppmv, averaged over 3 hours, and 60 ppmv averaged over 365-days. The second option would require measurement of TRS instead of H

2

S when any of the fuel gas burned in the process heater or other fuel gas combustion unit is from a coking unit. The TRS concentration limits would be the same as the H

2

S concentration limits. We are requesting comment on the proposed requirement to measure the TRS concentration. We are interested in any technological limitations of this option and whether there are other fuel gas streams that contain reduced sulfur compounds that should not be subject to the same requirement.

In addition to the proposed SO

2

emission limits and H

2

S and TRS concentration limits, we are also proposing to include the same exemptions from fuel gas continuous monitoring requirements that we are proposing for subpart J. See section IV.A of this preamble for a discussion of our rationale for these proposed exemptions.

NO

X

.

NO

X

emissions from process heaters are not subject to control under the existing NSPS in 40 CFR part 60, subpart J. However, several petroleum refiners are subject to NO

X

control requirements for process heaters in their consent decrees and State regulations. The emission limits to which refineries are subject vary from facility to facility. We evaluated four options as part of the BDT determination. Each option consists of a potential NO

X

emission limit and applicability based on process heater size. Option 1 would limit NO

X

emissions to 80 ppmv or less for all process heaters with a capacity greater than 20 million British thermal units per hour (MMBtu/hr). Option 2 would limit NO

X

emissions to 40 ppmv or less for all process heaters with a capacity greater than 20 MMBtu/hr. Option 3 would limit NO

X

emissions to 30 ppmv or less for all process heaters with a capacity greater than 40 MMBtu/hr. Option 4 would limit NO

X

emissions to 40 ppmv or less for process heaters with a capacity greater than 20 MMBtu/hr or less than or equal to 100 MMBtu/hr, and to 20 ppmv or less for process heaters with a capacity greater than 100 MMBtu/hr. In each option, the NO

X

concentration is based on a 24-hour rolling average.

The estimated fifth year emission reductions and costs for each option are summarized in Table 9. We believe that nearly all process heaters at refineries that will become subject to subpart Ja can meet Option 1 using combustion controls (low NO

X

burners or ultra low NO

X

burners). Stepping from Option 1 through Option 4 increases the fraction of process heaters that would need to use more efficient control technologies, such as LoTOx

TM

or SCR, to meet the NO

X

concentration limit. The options include a minimum 20 MMBtu/hr size threshold because none of the control technologies are cost effective for units with smaller capacities.

Table 9.—National Fifth Year Impacts Options for NO

X

Limits Considered for Process Heaters Subject to 40 CFR Part 60, Subpart Ja

Option

Total capital cost, $

(millions)

Total annual cost, $/yr

(millions)

Emission

reduction,

(tons

NO

X

/yr)

Cost effectiveness

($/ton)

Overall

Incremental

1

140

28

17,000

1,600

2

200

38

20,000

1,900

3,100

3

280

52

21,000

2,600

85,000

4

470

88

22,000

4,000

27,000

Based on the impacts in Table 9, the overall costs of option 1 and option 2 are reasonable compared to the emission reductions. The incremental cost, however, between options 1 and 2 is not commensurate with the additional emission reduction achieved. Therefore, BDT for process heaters greater than 20 MMBtu/hr was determined to be technology that achieves an outlet NO

X

concentration of 80 ppmv or less, and we are proposing standards for NO

X

emissions from process heaters consistent with this determination.

5. Work Practice Standards for Fuel Gas Production Units

We reviewed applicable state and local regulations and consent decree requirements and met with individual refinery representatives regarding their pollution prevention practices. The pollution prevention practices identified included flare minimization plans, fuel gas recovery requirements, start-up and shutdown requirements, and sulfur shedding plans (including redundant sulfur recovery capacity). Based on our review, all of these approaches could be expected to reduce emissions of VOC and SO

2

to the atmosphere. As described in the following subsections, we reviewed these pollution prevention practices and are proposing three different work practice standards. Work practice standards are being proposed because it is not feasible to prescribe or enforce a standard of performance for these emission sources. As provided in section 111(h) of the Clean Air Act, we may promulgate design, equipment, work practice, or operational standards when it is not feasible to prescribe or enforce a standard of performance. It is not feasible to prescribe or enforce a standard of performance for these

sources because either the pollution prevention measures eliminates the emission source, so that there are no emissions to capture and convey, or the emissions are so transient, and in some cases, occur so randomly, that the application of a measurement methodology to these sources is not technically and economically practical.

Elimination of Routine Flaring.

Flares are first and foremost a safety device used to reduce emissions from emergency pressure relief of gases from refinery process units. We in no way want to limit the use of flares for emergency releases. However, many refineries also routinely use flares as an emission control device under normal operating conditions.

Fuel gases produced within the refinery can be roughly divided into two categories based on the fuel gas stream pressure. Fuel gases produced in processes operated at higher pressures are easily routed to the fuel gas system; however, fuel gases that are produced from units operated near atmospheric pressures are not as easily routed to the fuel gas system. These “low pressure” fuel gases are often routed to flares because the flare gas system operates at a much lower pressure than the fuel gas system. Flare gas recovery systems are designed to compress the low pressure fuel gases, creating a high pressure fuel gas stream that can readily be added to the fuel gas system.

In 1998, the South Coast Air Quality Management District developed a rule requiring refineries to measure the flow rate and hydrocarbon content of the gases sent to a flare. This South Coast rule, although it did not set prescriptive emission limits, led to reduced flaring as refinery operators, armed with the monitoring results, identified cost-effective flare gas minimization or recovery projects. In 2005, South Coast amended this rule and established a no routine flaring goal based on the cost and anticipated emission reductions of flare gas recovery systems. The Bay Area Air Quality Management District also adopted a rule requiring flare monitoring in 2003 and adopted a rule to minimize flaring in 2006.

We considered adopting the South Coast and Bay Area rules for this NSPS for new flare systems. However, many refinery flares operate for 50 years, so very few flares or flare systems are expected to become subject to NSPS requirements, even after several decades. Instead, we are proposing to add “fuel gas producing units” as a new affected source under subpart Ja and focus the requirement on eliminating routine flaring of fuel gas at the process units producing the fuel gas. A refinery owner or operator installing a new process unit that produces low pressure fuel gas has options for eliminating routine flaring, including, but not limited to, diverting the fuel gas to a nearby low-pressure heater or boiler, pressurizing the fuel gas so that it can be diverted to the fuel gas system, or installing a flare gas recovery system. The proposed work practice standard is designed to allow flexibility in compliance approaches without imposing undue restrictions on the use of flares during malfunctions or other conditions wherein flaring is the best environmental management practice considering the safety of the plant personnel and surrounding people. Additionally, several new fuel gas producing units are expected to be installed every year, so by regulating the fuel gas producing units we not only provide flexibility, but we also increase the rate at which the no routine flaring requirement is implemented within the industry.

The impacts for this work practice are highly dependent on the amount of fuel gas generated by different fuel gas combustion units. Recovered fuel gas reduces the amount of natural gas a refinery must purchase to operate their process heaters. For example, fuel gases generated by fluid catalytic cracking units and coking units are routinely recovered into the fuel gas system due to the quantity of fuel gas generated in the process. For these systems, the savings associated with the recovered fuel gas provides a return on the capital investment associated with the compressor and ancillary equipment needed to recover the fuel gas. For other fuel gas producing units, such as reforming units, it is possible to route the fuel gas directly to the unit's process heater without additional gas compression. For a few refineries, a flare gas recovery system may be used.

We estimated planning and design costs for assessing methods to recover or otherwise avoid the release of fuel gas from new fuel gas producing units. As described previously, for many fuel gas producing units, the cost savings associated with the recovered fuel recovers the costs of the recovery equipment within the life-span on the equipment so that the annualized cost of controls is zero or slightly negative (indicating a cost savings). As a worst-case scenario, we used the impacts developed by the Bay Area for a system-wide flare gas recovery system. The total annualized cost of the system was estimated to be approximately $2 million; no credit was provided for the heating value of the flare gas recovered. VOC emission reductions were estimated to be approximately 1,000 tons per year and SO

2

emissions were estimated to be 3,500 tons per year. The cost-effectiveness on the flare gas recovery system was estimated to be approximately $2,000/ton of VOC removed and approximately $570/ton of SO

2

removed, assuming total costs are assigned to each pollutant. Therefore, even when fuel credits are not considered, flare gas recovery is cost-effective as an emissions control device. When properly sized, these flare gas recovery systems can eliminate all routine flaring. Therefore, eliminating routine flaring by use of fuel gas recovery, in-process fuel use, or system wide flare gas recovery is determined to be BDT.

We request comment on alternative means of eliminating routine flaring. As noted previously, a simple requirement to monitor gas flow and composition of gases sent to the flares resulted in reduced use of flares. An exemption from this monitoring requirement for flare systems that install flare gas recovery could provide refineries an incentive to install flare gas recovery systems. We request comment on this alternative and on the need to monitor flares that have flare gas recovery systems to ensure that the flare gas recovery system is properly sized and that no routine flaring is occurring.

Additionally, we understand that there are a limited number of refineries that produce more fuel gas than they can use in the refinery process heaters or steam boilers. These “fuel gas rich” refineries contend that flaring is BDT for these refineries. Although we believe that other options exist, such as building an electric co-generating unit, the cost-effectiveness of such an endeavor is very site-specific. We cannot conclude at this time that co-generation or other projects that use fuel gas are BDT. Therefore, we are co-proposing no requirement for fuel gas producing units. We request comment on the actual number and location of “fuel gas rich” refineries. We also request comment and data regarding the technical and economical feasibility of alternatives for “fuel gas rich” refineries to avoid routine flaring.

Emission Prevention During Start-up, Shutdown, and Malfunctions

. The current NSPS includes no requirements for a start-up, shutdown, and malfunction plan. We identified three emission prevention methods that can be addressed within the context of a start-up, shutdown, and malfunction plan. These are: Flare minimization during planned start-ups and shutdowns; flare minimization during malfunctions of the sour gas amine

treatment units and sulfur recovery plants; and performing root-cause analyses of malfunctions that release in excess of 500 lb per day of SO

2

. Our rationale for including each of these three emission prevention methods are described in the following paragraphs.

Flaring and direct venting of certain gas streams have been routinely used during planned start-up and shutdown of process units to quickly bring a process unit online or offline. These flaring and venting episodes have traditionally been exempt from any emission limitations. Nonetheless, some refineries have chosen to evaluate their start-up and shutdown emissions and alter their procedures so as to reduce or eliminate direct venting or flaring during planned start-up and shutdown events.

Typically, alternative start-up and shutdown procedures that reduce atmospheric emissions or flaring require more time to complete than conventional procedures. Therefore, there is a cost associated with the alternative procedures in terms of potential product/productivity loss. For refineries that have system-wide flare gas recovery systems, it may be a simple matter of scheduling the start-up or shutdown during a time when limited other flare gas is being generated so as to not overwhelm the flare gas recovery system. The cost-effectiveness of the alternative procedures would depend on the amount of gas flared or vented using the traditional procedures, the amount of these emissions that can be avoided using alternative procedures, the amount of product lost due to the increased start-up/shutdown time period, and the value of that product. As such, it is difficult to conclude that significant or complete emission reductions during planned start-up or shutdown events will be cost-effective under all conditions; therefore, we chose not to set a specific venting or flaring limit (or prohibition).

We estimate that the engineering review revision of a unit's start-up and shutdown plan would require approximately 20 engineering hours per process unit, at total cost of $1,300 to $1,500 per process unit (one-time costs). Assuming the unit requires maintenance shut-down only once every 5 years and the revised procedures only reduce VOC and SO

2

emissions by 1 ton each per event, the cost-effectiveness of the engineering review is $1,300 to $1,500 per ton of VOC and the same for SO

2

.

Based on this simplistic analysis, we are proposing that implementing a start-up and shutdown plan focused on reducing emissions during planned start-up and shutdown events would be BDT.

We evaluated several different requirements to promote continuous compliance with the SO

2

emission limits associated with fuel gas combustion devices and sulfur recovery plants even during times of process upsets or malfunctions associated with the amine system or sulfur recovery plant. “Process upset gas” is “gas

generated

by a petroleum refinery process unit as a result of upset or malfunction.” Process upset gas is exempt from the SO

2

emission limits. However, when there is a malfunction of the amine treatment system or the sulfur recovery plant, there has been some uncertainty as to whether combustion or flaring of the sour gas is considered to be exempt from the SO

2

emission limit. This is because the amine treatment system or sulfur recovery plant is not “generating” the gas stream, it is merely treating it. As such, the amine treatment system and sulfur recovery plant are essentially control devices, and refinery owners and operators are required to minimize emissions during these control system malfunctions, up to and including the shutdown of the emissions generating units.

A variety of prescriptive requirements were reviewed, such as requiring 24-hour storage capacity of lean amine solution and empty tank storage capacity to receive 24 hours worth of rich amine solution, requiring inventory of critical spare parts, and requiring redundant amine scrubbing and sulfur recovery capacity. While these are all viable options that a plant can employ to minimize malfunction emissions associated with the amine treatment system or sulfur recovery plant, the most cost-effective means to minimize these emissions are highly site-specific, being dependent on the number and location of the amine units or sulfur recovery trains within the sulfur recovery plant.

We evaluated two alternatives, which are not mutually exclusive, for minimizing flaring of H

2

S-rich fuel gas in the event of a malfunction in the amine stripper or sulfur recovery plant. Option 1 is to store 24 hours worth of lean amine solution in case of a malfunction in the amine stripper. We estimate that this alternative would require a capital cost of approximately $10 million (for 2 storage tanks and excess amine) for a 50 long LTD SRU system, resulting in an annualized cost of $1 million/year. If the 24 hours of excess amine was used one time per year for an entire day, 50 LTD of sulfur would have resulted in 110 tons of SO

2

emissions avoided. If there are three occurrences per year where the excess amine solution is used, 330 tons of emissions would be reduced. This scenario results in a cost-effectiveness ranging from $3,000 to 9,000 per ton of SO

2

reduced.

Option 2 is to have a redundant Claus unit. The capital cost of a 50 LTD Claus unit is also approximately $10 million, resulting in an annualized cost of $1 million/year. Again, if there are one to three days of emissions avoided, this option results in a cost-effectiveness ranging from $3,000 to $9,000 per ton of SO

2

reduced. For sulfur recovery plants consisting of multiple Claus units, the likelihood of needing the additional Claus train more than three times per year increases significantly, making the redundant Claus unit a cost-effective option.

It is difficult to predict the quantity of emissions avoided as they are dependent on random malfunction events of variable durations. While the cost-effectiveness values of these options are not necessarily compelling given the uncertainty in the emissions avoided, the options evaluated are expected to be extreme measures. It is likely, for example, that maintaining appropriate spare parts for the system would provide a cost-effective means of reducing emissions. This, along with short-term reductions in high-sulfur fuel gas production could be used to eliminate the need to flare or otherwise combust these high sulfur-containing fuel gases.

Based on this analysis, we are proposing that a start-up, shutdown, and malfunction plan that specifically addresses the minimization of fuel gas combustion of high sulfur-containing fuel gases during malfunctions of an amine treatment system or sulfur recovery plant is BDT. The start-up, shutdown, and malfunction plan will address specific process upset and malfunction events associated with the amine treatment system and sulfur recovery plant and the standard operating procedures to follow to minimize emissions during these events. Compliance is demonstrated by following the procedures in the plan. As previously mentioned, we are proposing a work practice standard rather than an equipment standard to provide flexibility to the refinery owner or operator regarding the best way to minimize malfunction emissions given the refinery's specific configuration and sulfur loads.

Finally, we evaluated a requirement for performing root-cause analyses as a means to minimize the frequency of process malfunctions and thereby

reduce malfunction emissions. Even though process upset gas is exempt from the SO

2

emission limits associated with fuel gas combustion units, we believe it is good air pollution practice to investigate the causes of significant atmospheric releases caused by process upsets or malfunctions to determine if similar upsets or malfunctions can be reasonably prevented from recurring. Similarly, we believe it is good pollution control practice to investigate significant emission exceedances to determine the cause of the exceedance and to implement procedures to prevent its recurrence. The cost-effectiveness of these investigations is dependent on the frequency and magnitude of the emission episodes; for very small emission episodes, the manpower required to perform the investigations do not justify the potential emission reductions that might be realized from the root-cause analysis. We estimate that a root-cause analysis would cost approximately $2,500 to perform. For emissions of less than 500 pounds per day, the cost-effectiveness of the root-cause analysis, even assuming it would completely eliminate a future recurrence, would be approximately $10,000 per ton of SO

2

reduced. Similarly, for emissions of 1,000 pounds per day, the cost-effectiveness would be on the order of $5,000 per ton of SO

2

reduced. As the probability of successfully identifying a means to avoid future emissions from each root-cause analysis performed is certainly less than 100 percent, we determined that it was not cost effective to perform root-cause analyses for SO

2

emissions exceedances of 500 pounds per day or less and request comment on alternative thresholds in the range of 500 to 1,000 lbs per day.

For SO

2

releases of greater than 500 pounds per day, the emissions reductions potential of the root-cause analyses increases and the cost-effectiveness improves, so we are proposing that performing root-cause analyses for SO

2

releases of greater than 500 pounds per day would be BDT. Any emission limit exceedance or any process start-up, shutdown, upset or malfunction that causes a discharge into the atmosphere in excess of 500 pounds per day of SO

2

would require a root cause analysis to be performed. We also considered a similar requirement for hydrocarbon flaring events with the purpose of reducing VOC emissions. However, we expect refinery owners and operators to investigate large hydrocarbon releases as these releases represent lost revenues. Furthermore, as flares are efficient in destroying VOC, the potential to significantly reduce VOC emissions by performing root-cause analysis is much less than the potential for reducing SO

2

emissions. We request comment on the need to include root cause analyses for hydrocarbon releases. If root-cause analyses are recommended, please provide in your comments the recommended release quantities that would trigger the root-cause analysis and justification for the recommendation. If root cause analyses are not recommended, please provide in your comments the rationale for not requiring root-cause analysis for any VOC (hydrocarbon) releases.

The proposed rule is intended to provide flexibility for each refinery owner and operator to develop procedures that are efficient and effective for their process configuration. The scope of these requirements is limited to affected facilities under this rule. We request comment on the need to implement this requirement to all new process units at the refinery, not just fuel gas producing units such as fluid catalytic cracking units, fluid coking units, fuel gas combustion devices, and sulfur recovery plants.

On the other hand, based on site-specific conditions and given the nature of the types of emissions events that are being addressed by the start-up, shutdown, and malfunction plan, it is impossible to conclusively determine that one or all of the emission reduction methods addressed in the start-up, shutdown, and malfunction plan will achieve any set level of emissions reduction or that those reductions, if any, will be cost-effective. Therefore, we are co-proposing no requirement for a start-up, shutdown, and malfunction plan. We request comments and supporting data that indicate the emission reductions that could be reasonably expected from a flare minimization plan for planned start-up and shutdown events, the number of planned events that occur per year (or over a 5 year period), and any other information that can be used to justify either the inclusion or exclusion of this provision in the final rule. We also request comments and supporting data that indicate the number and duration of malfunctions in the amine stripper and sulfur recovery plants, the costs associated with alternative sulfur shedding practices, and other information that can be used to justify either the inclusion or exclusion of this provision in the final rule.

Finally, we request comment, along with supporting data, that indicate the frequency of emission events exceeding 500 pounds per day, the percentage of times the root-cause analysis results in positive steps that may avoid future recurrence of the event, and other information that can be used to justify either the inclusion or exclusion of this provision in the final rule.

Delayed Coking Unit Depressurization.

The primary emission releases from delayed coking units occur as the coking vessels are depressurized and petroleum coke is removed from the unit. When the delayed coking cycle is completed, the coke-filled vessel is steam stripped. Most of the gases from this process continue to be sent to the coking unit distillation column. At some point in time, the steam gas discharge is diverted to the blow-down system. The delayed coking unit typically has a fuel gas recovery system (compressor) due to the quantity of fuel gas produced by the unit. Therefore, it is cost-effective to require the blow-down system gases to be recovered in the unit's fuel gas recovery system, in keeping with the proposed work practice standard that fuel gas from fuel gas producing units will not be routinely flared.

As the process unit continues to depressurize, there is a point where the gases can no longer be discharged to the blow-down system or fuel gas recovery line, at which point the remaining steam and gases are vented to the atmosphere. To achieve maximum reduction of uncontrolled releases, the unit should be depressurized to as low a pressure as possible before venting to the atmosphere. Below a pressure of 5 pounds per square inch gauge (psig) in the delayed coking unit drum, it is not technically feasible to divert the emissions for recovery. Above a vessel pressure of 5 psig, it is technically feasible to divert the emissions for recovery. Furthermore, as the unit already has a gas compressor, the costs associated with recovering these gases is minimal.

We estimate that this practice can reduce VOC emissions by 120 tons per year and SO

2

emissions by at 200 tons per year. The total annualized costs are expected to be minimal for new units, but installing the appropriate piping for a modified or reconstructed unit may result in annualized costs of up to $100,000 per year. Even under this extreme condition, the cost effectiveness of the requirement is about $800 per ton of VOC reduced and $500 per ton of SO

2

reduced. Therefore, we conclude that a work practice standard that requires a delayed coking unit to depressure to 5 psig during reactor vessel depressuring and vent the exhaust gases to the fuel gas system for recovery is BDT. Note this determination is independent of the

work practice to eliminate routine flaring from fuel gas producing units and requires flare gas recovery of depressurization gases even under the option of no work practice requirement to minimize flaring.

In addition to the depressurization emissions, we also identified at least one refinery that has designed an enclosed system for their coke-cutting operations. Coke cutting operations were identified as a significant VOC emission source at refineries during an Alberta Research Council study, with an estimated VOC emissions rate of 1,300 tons per year. We do not have any data regarding the effectiveness of the coke-cutting enclosure system, whether the enclosure seals are air tight or if they allow some percentage of the emissions escape. The enclosure may simply suppress the emissions until the coke is removed from the unit, at which time the emissions are released. Additionally, we do not have any data on the costs of these systems and whether or not existing units can be retrofitted if the delayed coking unit is modified or reconstructed. Therefore, we cannot conclude that an enclosed coke cutting system is BDT, but we request comment and additional information on coke-cutting system controls, their cost, their effectiveness, and their limitations.

VI. Modification and Reconstruction Provisions

Existing affected sources that are modified or reconstructed would be subject to the proposed standards in 40 CFR part 60, subpart Ja. A modification is any physical or operational change to an existing facility which results in an increase in the emission rate to the atmosphere of any pollutant to which a standard applies (see 40 CFR 60.14). Changes to an existing facility that do not result in an increase in the emission rate, as well as certain changes that have been exempted under the General Provisions (see 40 CFR 60.14(e)) are not considered modifications.

Rebuilt petroleum refinery process units would become subject to the proposed standards in 40 CFR part 60, subpart Ja under the reconstruction provisions, regardless of changes in emission rate. Reconstruction means the replacement of components of an existing facility such that (1) the fixed capital cost of the new components exceeds 50 percent of the fixed capital cost that would be required to construct a comparable entirely new facility; and (2) it is technologically and economically feasible to meet the applicable standards (40 CFR 60.15).

With the exception of the standards for fluid catalytic cracking units, we are proposing that modified or reconstructed sources be subject to the same proposed standards in 40 CFR part 60, subpart Ja, as new sources. The decision to maintain consistent standards for both new and modified or reconstructed sources was based on an evaluation of the cost-effectiveness and incremental cost-effectiveness of the proposed standards on both types of sources and on the feasibility of retrofitting existing units. We have included in the docket a table (Impacts Summary) which summarizes our estimates costs for different control options for both new and reconstructed or modified process units. We request comment on these cost estimates and on specific issues related to the feasibility of retrofitting existing units, as well as our assessment that cost-effectiveness numbers are similar enough such that it is appropriate to have identical standards for both new and modified or reconstructed sources.

VII. Request for Comments

Table 10 summarizes the topics on which we have specifically requested comment throughout this preamble. We note, however, that comments on all aspects of this proposal are welcome.

Table 10.—Summary of Topics on Which Comment Is Requested

Topic

Section in this

preamble where

topic is discussed

Effects of proposed PM, SO

2

and NO

X

standard on modified or reconstructed fluid catalytic cracking units. Also co-proposed 40 CFR part 60, Subpart J standards for SO

2

and PM and no NO

X

limits for modified and reconstructed sources

III.B. and V.E.1.

Exemption for emergency flares

IV.A.

Exemption from monitoring for fuel gas streams related to commercial liquid products

IV.A.

Exemption from monitoring for fuel gas streams generated by process units that are intolerant of sulfur

IV.A.

Alternative PM limit for fluid catalytic cracking units based on condensable PM as well as filterable PM

V.E.1.

Alternative lower (20 ppmv, 40 ppmv) NO

X

limit, averaged over 365 days, for fluid catalytic cracking units

V.E.1.

Co-propose no new NO

X

standard for fluid coking units

V.E.2.

Appropriate long-term average H

2

S concentration limit for fuel gas combustion units, and requirement to monitor TRS instead of H

2

S for fuel gas from coker units

V.E.4.

Various aspects of work practice standards to minimize routine flaring and enhance SO

2

control versus no standards: alternative means of eliminating flaring, number of “fuel gas rich” refineries, need for a startup, shutdown, and malfunction plan (SSMP), including rationale for or against requiring a root cause analysis for hydrocarbon releases and sulfur shedding practices, and information about emission control systems for coke cutting operations. Also co-propose no requirements for routine flaring and no SSMP

V.D.5.

VIII. Summary of Cost, Environmental, Energy, and Economic Impacts

In setting standards, the CAA requires us to consider alternative emission control approaches, taking into account the estimated costs as well as impacts on energy, solid waste, and other effects. We request comment on whether we have identified the appropriate alternatives and whether the proposed standards adequately take into consideration the incremental effects in terms of emission reductions, energy, and other effects of these alternatives. We will consider the available information in developing the final rule.

A. What are the impacts for petroleum refining process units?

We are presenting estimates of the impacts for the proposed requirements of subpart Ja that change the performance standards: the emission limits for fluid catalytic cracking units, sulfur recovery plants, fluid coking units, fuel gas combustion devices, and process heaters, as well as the work practice standards. The proposed amendments to 40 CFR part 60, subpart J are clarifications to the existing rule, and they have no emission reduction impacts. The cost, environmental, and economic impacts presented in this section are expressed as incremental

differences between the impacts of petroleum refining process units complying with the proposed subpart Ja and the current NSPS requirements of subpart J (i.e., baseline). The impacts are presented for petroleum refining process units that commence construction, reconstruction, or modification over the next 5 years. The analyses and the documents referenced below can be found in Docket ID No. EPA-HQ-OAR-2007-0011.

In order to determine the incremental costs and emission reductions of this proposed rule, we first estimated baseline impacts. For new sources, baseline costs and emission reductions were estimated for complying with subpart J; incremental impacts for subpart Ja were estimated as the costs to comply with subpart J subtracted from the costs to comply with proposed subpart Ja. Sources that are modified or reconstructed over the next 5 years would comply with subpart J in the absence of proposed subpart Ja. We assumed that prior to reconstruction or modification, these sources would either be subject to a consent decree (equivalent to about 77 percent of the industry by capacity), complying with subpart J or equivalent limits, or complying with 40 CFR part 63, subpart UUU (MACT II). Baseline costs and emission reductions were estimated as the effort needed to comply with subpart J from one of those three starting points. The costs and emission reductions to comply with proposed subpart Ja were estimated from those starting points as well. The estimated costs presented for work practice standards include only the labor cost to prepare the required plan or analysis; we did not attempt to quantify costs and emission reductions for the variety of ways a facility may choose to implement those plans. We assumed that each facility would evaluate their options and choose the most cost-effective option for the facility's unique position. For further detail on the methodology of these calculations, see Docket ID No. EPA-HQ-OAR-2007-0011.

When considering and selecting emission limits for the proposed rule, we evaluated the cost-effectiveness of each option for new sources separately from reconstructed and modified sources. However, since our selections for each process unit and pollutant were consistent for all units, we are presenting our costs and emission reductions for the overall rule. We estimate that the proposed amendments will reduce combined emissions of PM, SO

2

, and NO

X

about 55,800 tons/yr from the baseline. The estimated increase in annual cost, including annualized capital costs, is about $54,100,000. The overall cost-effectiveness is about $970 per ton of pollutants removed. The estimated nationwide 5-year incremental emissions reductions and cost impacts for the proposed amendments are summarized in Table 11 of this preamble.

Table 11.—National Incremental Emission Reductions and Cost Impacts for Petroleum Refinery Units Subject to Proposed Standards Under 40 CFR Part 60, Subpart Ja (Fifth Year After Proposal)

Process unit

Pollutant

Total capital cost

($1,000)

Total annual cost

($1,000/yr)

Annual

emission

reductions

(tons/yr)

Cost-

effectiveness

($/ton)

FCCU

PM and SO

2

40,000

9,500

9,500

1,000

FCCU

NO

X

28,000

7,300

3,500

2,100

Fluid Coker

PM and SO

2

14,000

4,800

23,000

210

Fluid Coker

NO

X

4,500

970

760

1,300

SRP

SO

2

1,100

680

550

1,200

Process Heaters and Fuel Gas Combustion

SO

2

0

2,880

1,300

2,200

Process Heaters

NO

X

140,000

28,000

17,000

1,600

Work Practices

250

Total

230,000

54,000

56,000

970

B. What are the secondary impacts?

Indirect or secondary air quality impacts of this proposed rule would result from the increased electricity usage associated with the operation of control devices. Assuming that plants would purchase electricity from a power plant, we estimate that the standards as proposed would increase secondary emissions of criteria pollutants, including PM, SO

2

, NO

X

, and CO from power plants. For new, modified or reconstructed sources, this proposed rule would increase secondary PM emissions by 24 Mg/yr (27 tpy); secondary SO

2

emissions by about 970 Mg/yr (1,100 tpy); secondary NO

X

emissions by about 480 Mg/yr (530 tpy); and secondary CO emissions by about 16 Mg/yr (17 tpy) for the 5 years following proposal.

As explained earlier, we expect that affected facilities will control emissions from fluid catalytic cracking units by installing and operating ESP or wet gas scrubbers. We also expect that the emissions from the affected fluid coker will be controlled with a wet scrubber. For these process units, we estimated solid waste impacts for both types of control devices and water impacts for wet gas scrubbers. In addition, the controls needed by small sulfur recovery plants will generate condensate. We project that this proposed rule will generate 4.5 billion gallons of water per year for the 5 years following proposal. We also estimate that this proposed rule will generate 8,600 Mg/yr (7,800 tpy) of solid waste over those 5 years.

Energy impacts consist of the electricity and steam needed to operate control devices and other equipment that would be required under the proposed rule. Our estimate of the increased energy demand includes the electricity needed to produce the required amounts of steam as well as direct electricity demand. We project that this proposed rule would increase overall energy demand by about 170 gigawatt-hours per year (590 billion British thermal units per year).

C. What are the economic impacts?

This proposal affects certain new and reconstructed/modified sources found at petroleum refineries as defined earlier in this preamble. We performed an economic impact analysis that estimates changes in prices and output for gasoline nationally using the annual compliance costs estimated for this proposal. The methodology for this

analysis incorporates changes in producer and consumer behavior by considering passthrough of increased production costs from producers to consumers. All estimates are for the fifth year after proposal since this is the year for which the compliance cost impacts are estimated.

The analysis estimates a price increase in gasoline of less than 0.02 percent nationally will take place along with a corresponding reduction in gasoline output of less than 0.004 percent (or less than 6 million gallons a year). The overall total annual social costs, which reflect changes in consumer and producer behavior in response to the compliance costs, are $53.0 million (2005 dollars) or almost identical to the compliance costs.

For more information, please refer to the economic impact analysis report that is in the public docket for this proposed rule.

D. What are the benefits?

We estimate the monetized benefits of this proposed rule to be $957 million (2005$) in the fifth year after proposal. We base the portion of the benefits estimate derived from the PM

2.5

and SO

2

emission reductions on the approach and methodology laid out in EPA's 2004 benefits analysis supporting the regulation of emissions from the Industrial Boilers MACT (included in the Regulatory Impact Analysis (RIA) for the Industrial Boilers and Process Heaters NESHAP, February 2004). We chose the benefit analysis contained in this RIA as the basis for estimating the benefits from emission reductions of these two pollutants since most of the elements in that rule are similar to those covered here. These elements, which are the stack height, a number of the controls applied, and the pollutants affected—PM

2.5

and SO

2

, but not NO

X

—are similar to those covered by the Industrial Boiler MACT standard.

We base the portion of the benefits estimate derived from the NO

X

emission reductions on the approach and methodology laid out in EPA's 2005 benefits analysis supporting the regulation of emissions from the Clean Air Interstate Rule (CAIR) (included in the Regulatory Impact Analysis for the Clean Air Interstate Rule, March 2005). We chose the CAIR analysis as the basis for estimating the benefits from emission reductions of this pollutant since most of the elements in CAIR are similar to those covered here. These elements, which are the stack height, a number of the controls applied, and the pollutant affected—in this case, NO

X

only—are similar to those covered by CAIR. These three factors lead us to believe that we might reasonably estimate benefits for this proposed rule using a benefits transfer approach and values from the Industrial Boilers MACT analysis for estimating the SO

2

and PM

2.5

benefits of this rule, and the CAIR analysis for the NO

X

benefits of the rule. Specifically, these estimates are based on application of the benefits scaling approach derived from the benefits analyses completed for these rulemakings. As mentioned above, the methodologies are laid out in the Industrial Boilers MACT and CAIR RIA. A summary of the benefits estimates is in Table 12 below.

1

1

We use the SO

2

benefits/ton estimate derived from the Industrial Boilers MACT benefit analysis based on the factors listed above. We also note that the SO

2

benefits/ton estimate derived from the CAIR benefits analysis is $18,000 in 2010 and $22,000 in 2015, both of which are quite close to the estimate we use in this analysis. We use the NO

X

benefits/ton estimate from the CAIR Boilers MACT benefits analysis (no NO

X

reductions take place as a result of the Industrial Boilers MACT).

Table 12.—Summary of Benefits Estimates For Proposed NSPS

Pollutant

Monetized benefits per ton emission reduction

Emission

reductions

(tons)

Total monetized benefits*

(millions of 2005 dollars)

PM

2.5

$88,000

3,221

$283.4

SO

2

20,000

31,358

627.2

NO

X

2,200

21,266

46.8

Grand Total:

$957.4

* All estimates are for the analysis year (fifth year after proposal). Emission reductions reflect the combination of proposed options for both new and reconstructed/modified sources.

The specific estimates of benefits per ton of pollutant reductions included in this analysis are largely driven by the concentration response function for premature mortality, which is based on the American Cancer Society cohort (ACS) (Pope, C.A. III, et al., “Lung Cancer, Cardiopulmonary Mortality, and Long-Term Exposure to Fine Particulate Air Pollution,” JAMA, 2002). Since the publication of CAIR, the EPA's Office of Air and Radiation has adopted a different format for its benefits analysis in which characterization of the uncertainty in the concentration response function is integrated into the main benefits analysis. The PM NAAQS analysis provides an indication of the sensitivity of our results to the use of alternative concentration response functions, including those derived from the recently completed expert elicitation study. Specifically, compared to the final PM NAAQS estimate of the mean mortality from the ACS cohort, the expert-based premature mortality incidence ranged from 50 percent of the mean ACS estimate to more than five times the size of the ACS mean estimate. The Agency is currently updating the estimates used here to calculate the benefits of the proposed NSPS and intends to consider using these updated benefits estimates as part of an approach similar to that used in the PM NAAQS RIA in the benefits analyses for the final NSPS.

With the annualized costs of this rulemaking estimated at $54 million (2005$) in the fifth year after proposal and with estimated benefits of $957 million (2005$) for that same year, EPA believes that the benefits are likely to exceed the costs by a significant margin even when taking into account the uncertainties in the cost and benefit estimates. For more information, please refer to the RIA for this proposed rule that is available in the docket.

IX. Statutory and Executive Order Reviews

A. Executive Order 12866: Regulatory Planning and Review

Under section 3(f)(1) of Executive Order 12866 (58 FR 51735, October 4, 1993), this action is an “economically significant regulatory action” because it is likely to have an annual effect on the

economy of $100 million or more. Accordingly, EPA submitted this action to the Office of Management and Budget (OMB) for review under Executive Order 12866 and any changes made in response to OMB recommendations have been documented in the docket for this action.

In addition, EPA prepared an analysis of the potential costs and benefits associated with this action. This analysis is contained in the Regulatory Impact Analysis (RIA) for the Proposed Petroleum Refinery NSPS, EPA-452/R-07-006. A copy of the analysis is available in the docket for this action and the analysis is briefly s

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