EPA 452/R-08-002
April 2008
Regulatory Impact Analysis of the Petroleum Refinery NSPS
By:
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, North Carolina
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Benefits and Cost Group
Research Triangle Park, NC
-------
Regulatory Impact Analysis of the Petroleum
Refinery NSPS
-------
SECTION 1
EXECUTIVE SUMMARY
EPA has characterized the facilities and companies potentially affected by the NSPS by
examining existing refineries and the companies that own them. EPA projects that new refineries
and processes will be similar to existing ones, and that the companies owning new sources will
also be similar to the companies owning existing refineries. EPA has collected data on 150
existing refineries, owned by 58 companies. Of the affected parent companies, twenty-five are
identified as small entities based on the Small Business Administration size standard criteria for
NAICS 324110, for they employ 1,500 or fewer employees.
EPA estimates that complying with the final NSPS will have an annualized cost of
approximately $31 million per year (2006 dollars) in the fifth year after proposal. Using these
costs, EPA estimates that the NSPS will have limited impacts on the market for motor gasoline.
Based on sales data obtained for the affected small entities, EPA estimates that, due to the
expected annualized cost savings, that the NSPS will not result in a SISNOSE (a significant
economic impacts for a substantial number of small entities).
The final petroleum refineries NSPS is considered subject to the requirements of Circular
A-4 because EPA expects that the sum of benefits and costs are potentially $1 billion or higher.
EPA's estimate of the benefits of the NSPS, based on information from the PM2.5 expert
elicitation study released in October, 2006, is a range from $220 million to $ 1.9 billion (2006
dollars) in the fifth year after proposal. EPA believes that the benefits are likely to exceed the
costs by a substantial margin under this rulemaking even when taking into account uncertainties
in the cost and benefit estimates.
1-1
-------
SECTION 2
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA's) Office of Air Quality Planning and
Standards (OAQPS) is currently revising the existing Subpart J New Source Performance
Standards (NSPS) for petroleum refineries. In addition, the Agency is adding a Subpart Ja that
provides new requirements, including new emissions limits for new fluid catalytic cracking units
(FCCUs), new fluid coking units, and new flares. This final regulation include emissions limits
for new and modified/reconstructed sources, and these limits are set for sulfur dioxide (SO2),
nitrogen oxides (NOx), coarse particulate matter (PM10), volatile organic compounds (VOC), and
other pollutants. This regulatory impact analysis (R1A), prepared in response to requirements
under Executive Order 12866, presents the results of analyses undertaken in support of this final
rule including compliance costs, benefits, economic impacts, and impacts to small businesses.
2.1 Introduction
The petroleum refining industry comprises establishments primarily engaged in refining
crude petroleum into refined petroleum. Examples of refined petroleum products include
gasoline, kerosene, asphalt, lubricants, solvents, and a variety of other products. Petroleum
refining falls under the North American Industrial Classification System (NAICS) 324110.
This RIA is organized as follows:
¦ Section 1: Executive Summary,
¦ Section 2: Introduction,
¦ Section 3 and an Appendix A: Profile of the Petroleum Refinery Industry,
¦ Section 4: NSPS Regulatory Alternatives, and Costs and Emission Reductions From
Complying with the NSPS, and an Appendix B: Summary of Significant Comments
and Responses, And Rationales for NSPS Emission Limits
¦ Section 5: Economic Impacts of the NSPS,
¦ Section 6: Potential Impacts on Small Businesses,
¦ Section 7: Benefits of the NSPS, and
¦ Appendix C: Overview of Economic Model Equations
2.2 Reason for Today's Action: Market Failure or Other Social Purpose
il
The petroleum nbfinery NSPS is of sufficient impact to be considered as falling under the
requirements.of Circular A-4, an addendum to the existing requirements for Executive Order
2-1
-------
12866 (OMB, 2003). This final regulation is being issued in response to a court-ordered
settlement between U.S. EPA and various parties requiring review of the existing NSPS.
As discussed in Circular A-4, among the reasons a regulation such as this one may also
be issued is to address market failure. The major types of market failure include: externality,
market power, and inadequate or asymmetric information. Correcting market failures is a reason
for regulation, but it is not the only reason. Other possible justifications include improving the
functioning of government, removing distributional unfairness, or promoting privacy and
personal freedom.
2.2.1 Externality, Common Property Resource, and Public Good
An externality occurs when one party's actions impose uncompensated benefits or costs
on another party. Environmental problems are a classic case of externality. For example, the .
smoke from a factory may adversely affect the health of local residents while soiling the property
in nearby neighborhoods. If bargaining were costless and all property rights were well defined,
people would eliminate externalities through bargaining without the need for government
regulation. From this perspective, externalities arise from high transactions costs and/or poorly
defined property rights that prevent people from reaching efficient outcomes through market
transactions.
Resources that may become congested or overused, such as fisheries or the broadcast
spectrum, represent common property resources. "Public goods," such as defense or basic
scientific research, are goods where provision of the good to some individuals cannot occur
without providing the same level of benefits free of charge to other individuals.
2.2.2 Market Power
Firms exercise market power when they reduce output below what would be offered in a
competitive industry in order to obtain higher prices. They may exercise market power
collectively or unilaterally. Government action can be a source of market power, such as when
regulatory actions exclude low-cost imports. Generally, regulations that increase market power
for selected entities should be avoided. However, there are some circumstances in which
government may choose to validate a monopoly. If a market can be served at lowest cost only
when production is limited to a single producer B local gas and electricity distribution services,
for example B a natural monopoly is said to exist. In such cases, the government may choose to
approve the monopoly and to regulate its prices and/or production decisions. Nevertheless, you
should keep in mind that technological advances often affect economies of scale. This can, in
2-2
-------
turn, transform what was once considered a natural monopoly into a market where competition
can flourish.
2.2.3 Inadequate or Asymmetric Information
Market failures may also result from inadequate or asymmetric information. Because
information, like other goods, is costly to produce and disseminate, your evaluation will need to
do more than demonstrate thepossible existence of incomplete or asymmetric information. Even
though the market may supply less than the full amount of information, the amount it does
supply may be reasonably adequate and therefore not require government regulation. Sellers
have an incentive to provide information through advertising that can increase sales by,
highlighting distinctive characteristics of their products. Buyers may also obtain reasonably
adequate information about product characteristics through other channels, such as a seller
offering a warranty or a third party providing information.
Even when adequate information is available, people can make mistakes by processing it
poorly. Poor information-processing often occurs in cases of low probability, high-consequence
events, but it is not limited to such situations. For instance, people sometimes rely on mental
rules-of-thumb that produce errors. If they have a clear mental image of an incident which makes
it cognitively "available," they might overstate the probability that it will occur. Individuals
sometimes process information in a biased manner, by being too optimistic or pessimistic,
without taking sufficient account of the fact that the outcome is exceedingly unlikely to occur.
When mistakes in information processing occur, markets may overreact. When it is time-
consuming or costly for consumers to evaluate complex information about products or services
(e.g., medical therapies), they may expect government to ensure that minimum quality standards
are met. However, the mere possibility of poor information processing is not enough to justify
regulation. If you think there is a problem of information processing that needs to be addressed,
it should be carefully documented.
2.2.4 Other Social Purposes
There are justifications for regulations in addition to correcting market failures. A
regulation may be appropriate when you have a clearly identified measure that can make
government operate more efficiently. In addition, Congress establishes some regulatory
programs to redistribute resources to select groups. Such regulations should be examined to
ensure that they are both effective and cost-effective. Congress also authorizes some regulations
to prohibit discrimination that conflicts with generally accepted norms within our society.
2-3
-------
Rulemaking may also be appropriate to protect privacy, permit more personal freedom or
promote other democratic aspirations.
2.3 References
U.S. Office of Management and Budget. Circular A-4, September 17, 2003. Found.on the
Internet at .
2-4
-------
SECTION 3
INDUSTRY PROFILE
3.1 Introduction
The U.S. Environmental Protection Agency's (EPA's) Office of Air Quality Planning and
Standards (OAQPS) is currently revising the existing Subpart J New Source Performance
Standards (NSPS) for petroleum refineries. In addition, the Agency is adding a Subpart Ja that
provides new requirements, including new emissions limits for new and modified and
reconstructed fluid catalytic cracking units (FCCUs), new fluid coking units, new process
heaters, and new flares. These standards include emissions limits for reductions of sulfur dioxide
(SO2), nitrogen oxides (NOx), coarse particulate matter (PM10), volatile organic compounds,
(
(VOC) and other pollutants. This industry profile of the petroleum refining industry provides
information that will support subsequent regulatory impact analyses (RIAs) and economic
impact analyses (ElAs) that will assess the impacts of these standards.
At its core, the petroleum refining industry comprises establishments primarily engaged
in refining crude petroleum into finished petroleum products. Examples of these petroleum
products include gasoline, kerosene, asphalt, lubricants, and solvents, among others.
Firms engaged in petroleum refining are categorized under the North American Industry
Classification System (NAICS) code 324110. In 2006, 149 establishments owned by 58 parent
companies were refining petroleum. That same year, the petroleum refining industry shipped
products valued at over $489 billion (U.S. Department of Commerce, Bureau of the Census,
2007).
This industry profile report is organized as follows. Section 3.2 provides a detailed
description of the inputs, outputs, and processes involved in petroleum refining. Section 3.3
describes the applications and users of finished petroleum products. Section 3.4 discusses the
organization of the industry and provides facility- and company-level data. In addition, small
businesses are reported separately for use in evaluating the impact on small business to meet the
requirements of the Small Business Regulatory Enforcement and Fairness Act (SBREFA).
Section 3.5 contains market-level data on prices and quantities and discusses trends and
projections for the industry.
3-1
-------
3.2 The Supply Side
Estimating the economic impacts of any regulation on the petroleum refining industry
requires a good understanding of how finished petroleum products are produced (the "supply
side" of finished petroleum product markets). This section describes the production process used
to manufacture these products as well as the inputs, outputs, and by-products involved. The
section concludes with a description of costs involved with the production process.
3.2.1 Production Process, Inputs, and Outputs
Petroleum pumped directly out of the ground, known as crude oil, is a complex mixture
of hydrocarbons (chemical compounds that consist solely of hydrogen and carbon) and various
impurities such as salt. To manufacture the variety of petroleum products recognized in every
day life, this tar-like mixture must be refined and processed over several stages. This section
describes the typical stages involved in this process as well as the inputs and outputs.
3.2.1.1 The Production Process
The process of refining crude oil into useful petroleum products can be separated into two
phases and a number of supporting operations. These phases are described in detail in the
following section. In the first phase, crude oil is desalted and then separated into its various
hydrocarbon components (known as "fractions"). These fractions include gasoline, kerosene,
naphtha, and other products (EPA, 1995).
In the second phase, the distilled fractions are converted into petroleum products (such as
gasoline and kerosene) using three different types of downstream processes: combining,
breaking, and reshaping (EPA, 1995). An outline of the refining process is presented in
Figure 3-1.
Desalting. Before separation into fractions, crude oil is treated to remove salts,
suspended solids, and other impurities that could clog or corrode the downstream equipment.
This process, known as "desalting," is typically done by first heating the crude oil, mixing it with
process water, and depositing it into a gravity settler tank. Gradually, the salts present in the oil
will be dissolved into the process water (EPA, 1995). After this takes place, the process water is
separated from the oil by adding demulsifier chemicals (a process known as chemical separation)
and/or by applying an electric field to concentrate the suspended water globules at the bottom of
the settler tank (a process known as electrostatic separation). The effluent water is then removed
from the tank and sent to the refinery wastewater treatment facilities (EPA, 1995). This process
is illustrated in Figure 3-2.
3-2
-------
Atmospheric Distillation. The desalted crude oil is then heated in a furnace to 750°F and
fed into a vertical distillation column at atmospheric pressure. After entering the tower, the
lighter fractions flash into vapor and travels up the tower. This leaves only the heaviest fractions
(which have a much higher boiling point) at the bottom of the tower. These fractions include
heavy fuel oil and asphalt residue (EPA, 1995).
r~
Crude *il (01
I
T
GAS
»EPARATtOt
i IF':-!yrricnc»ion
_J bJd 191
!-{> GASPLANT
POblMERIZATIOr' P'olyrocnsiiori
H\DF.00ESULfUR'
Lrght crude oil
digrtillate [2) .
,~A I lAIU'btton
I fc
Light SR naphtha (3)
CATALYTIC
ISOMERIZATlOn
> ALKYLATION
| naphtha } 10|
Ijc-naphtoh |_14i
Hewy SR naphth
.w
HYDRODESUirU-
ATMOSPHtRIC
DISTILLATION
Lt LR naphtha (3)
RIZATION'TREATING
1 i
A
Debited
crude oil (1)
SR Kerosene (5)
SR Middle di^ulht'
CATALYTIC
REFORMING
Rtn>:mv-. 115
Lt hydrocrached
r«)
SR Ga; oil (i
Atmorpheric
tO'A'tf
18)
i—£>1
ca-stim:N-
^|Hydrodc^ylru
U vacuum dirtilh*e [1'-IJ |
VACUUM
^ILLATION
)
•/acuun
tower
¦iridic
W)
Hvy vacuum
(20)
solved
DCA^PHALTIMij
CATALYTIC
CRA* KING
T
r.^phths(1«)
,t ot crocked >.
napMhi (22i
HD-? hvy hjphtha [A>
GASOLINE
(NAPHTHA)
?WEETENIH<3,
TREATING
AND
BLENDING
—F'j-1 "pai-i:
~t>- Liquihod
pctroteom qw ILF'G)
- Aviitior*
qasolifni
- Automotive
qacolific-
SF" heroine 15[
SR mid dirtillvc |6)
HDi mid distillate (6A1
LtT-jtcrachod disHllatt (£4)
DISTILLATE
SWEETENING
TREATING '
AMD
ELEHDiNG
Hvy vacuum distillate- {20}
Hvy >:g» mcl.^d dtshllatc- (26).
U thermal cracked distillate (oQlfGas oiQ
sphal'
VIS&P.EAKING
Cat ciacUd
clvmtd oil 12"
Thermally cracl.cd
Ii£KkKl21J^
Vacuum residue 1211
Lube Kcdsto-ih jiO
HVOROTREATINGr
SOLVENT
EXTRACTION
PaffuntcjSJ.
Atmvrphonc».-v?.r rtn (?)
FsESIOUAL
TREATING
AND
ELEHDING
SOLVENT
DEWAKtNG
HYDRO-
TREATING
AMD
BLENDING
Ge.vaxcdoil ^
|Rvffi',atc)
Ceoiled 'yy* ^
- Jet tuci;
- ktrosen?
- tolYcr>*s
- Distillito
ruil oil:-
- Diesel fue
oils
• Residual
fuel cilr
-lubricantr
-Grease
Figure 3-1. Outline of the Refining Process
Source: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23,2006.
3-3
-------
Process
water
Electrical
power
Q
Alternate
Alternate
Desalted
crude
—>
Unrefined
crude
Heater Emulsifier
Effluent
water
»
Figure 3-2. Desalting Process
Source: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23, 2006.
As the hot vapor rises, its temperature is gradually reduced. Lighter fractions condense onto
trays located at successively higher portions of the tower. For example, motor gasoline will
condense at higher portion of the tower than kerosene because it condenses at lower temperatures.
This process is illustrated in Figure 3-3. As these fractions condense, they will be drawn off their
respective trays and potentially sent downstream for further processing (OSHA, 2003; EPA, 1995).
Figure 3-3. Atmospheric Distillation Process
Source: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23, 2006.
" od
Purrip
3-4
-------
Vacuum Distillation. The atmospheric distillation tower cannot distil the heaviest
fractions (those at the bottom of the tower) without cracking under requisite heat and pressure.
So these fractions are separated using a process called vacuum distillation. This process takes
place in one or more vacuum distillation towers and is similar to the atmospheric distillation
process, except very low pressures are used to increase volatization and separation. A typical
first-phase vacuum tower may produce gas oils or lubricating-oil base stocks (EPA, 1995). This
process is illustrated in Figure 3-4.
Downstream Processing. To produce the petroleum products desired by the market
place, most fractions must be further refined after distillation or "downstream." These
downstream processes change the molecular structure of the hydrocarbon molecules by breaking
them into smaller molecules, joining them to form larger molecules, or shaping them into higher
quality molecules (EPA, 1995).
To vacuum system
Vacuum
Residuum ~-
VWV
gas oil
lubricating oils
Vacuum
residuum
Furnace
Figure 3-4. Vacuum Distillation Process
Source: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23, 2006.
Downstream processes include thermal cracking, coking, catalytic cracking, catalytic
hydrocracking, hydrotreating, alkylation, isomerization, polymerization, catalytic reforming,
solvent extraction, merox, dewaxing, propane deasphalting and other operations (EPA, 1995).
3-5
-------
3.2.1.2 Supporting Operations
In addition to the processes described above, there are other refinery operations that do
not directly involve the production of hydrocarbon fuels, but serve in a supporting role. Some of
the major supporting operations are described in this section.
Wastewater Treatment. Petroleum refining operations produce a variety of wastewaters
including process water (water used in process operations like desalting), cooling water (water
used for cooling that does not come into direct contact with the oil), and surface water runoff
(resulting from spills to the surface or leaks in the equipment that have collected in drains).
Wastewater typically contains a variety of contaminants (such as hydrocarbons,
suspended solids, phenols, ammonia, sulfides, and other compounds) and must be treated before
it is recycled back into refining operations or discharged. Petroleum refineries typically utilize
two stages of wastewater treatment. In primary wastewater treatments, oil and solids present in
the wastewater are removed. After this is completed, wastewater can be discharged to a publicly
owned treatment facility or undergo secondary treatment before being discharged directly to
surface water. In secondary treatment, microorganisms are used to dissolve oil and other organic
pollutants that are present in the wastewater (EPA, 1995; OSHA, 2003).
Gas Treatment and Sulfur Recovery. Petroleum refinery operations such as coking and
catalytic cracking emit gases with a high concentration of hydrogen sulfide mixed with light
refinery fuel gases (such as methane and ethane). Sulfur must be removed from these gases in
order to comply with Clean Air Act's SOx emission limits and to recover saleable elemental
sulfur.
Sulfur is recovered by first separating the fuel gases from the hydrogen sulfide gas. Once
this is done, elemental sulfur is removed from the hydrogen sulfide gas using a recovery system
known as the Claus Process. In this process, hydrogen sulfide is burned under controlled
conditions producing sulfur dioxide. A bauxite catalyst is then used to react with the sulfur
dioxide and the unburned hydrogen sulfide to produce elemental sulfur. However, the Claus
process only removed 90% of the hydrogen sulfide present in the gas stream, so other processes
must be used to recover the remaining sulfur (EPA, 1995).
Additive Production. A variety of chemicals are added to petroleum products to
improve their quality or add special characteristics. For example, ethers have been added to
gasoline to increase octane levels and reduce CO emissions since the 1970s.
3-6
-------
The most common ether additives being used today are methyl tertiary butyl ether
(MTBE), and tertiary amyl methyl ether (TAME). Larger refineries tend tomanufacture these
additives themselves by reacting isobutylene (a by-product of several refinery processes) with
methanol (OSHA, 2003).
Heat Exchangers, Coolers, and Process Heaters. Petroleum refineries require very
high temperatures to perform many of their refining processes. To achieve these temperatures,
refineries use fired heaters fueled by refinery or natural gas, distillate, and residual oils. This heat
is managed through heat exchanges, where are composed of bundles of pipes, tubes, plate coils,
and other equipment that surround heating or cooling water, steam, or oil. Heat exchanges
facilitate the indirect transfer of heat as needed (OSHA, 2003).
Pressure Release and Flare Systems. As liquids and gases expand and contract through
the refining process, pressure must be actively managed to avoid accident. Pressure-relief
systems enable the safe handling of liquids and gases that that are released by pressure-relieving
devices and blow-downs. According to the OSHA Technical Manual, "pressure relief is an
automatic, planned release when operating pressure reaches a predetermined level. A blow-down
normally refers to the intentional release of material, such as blow-downs from process unit
startups, furnace blow-downs, shutdowns, and emergencies" (OSHA, 2003).
Blending. Blending is the final operation in petroleum refining. It is the physical mixture
of a number of different liquid hydrocarbons to produce final petroleum products that have
desired characteristics. For example, additives such as ethers can be blended with motor gasoline
to boost performance and reduce emissions. Products can be blended in-line through a manifold
system, or batch blended in tanks and vessels (OSHA, 2003).
3.2.1.3 Inputs
The inputs in the production process of petroleum products include general inputs such as
labor, capital, and water. The inputs specific to this industry are crude oil and the variety of
chemicals used in producing petroleum products. These two specific inputs are discussed below.
Crude Oil. Contrary to popular conception, crude oils are complex, heterogeneous
mixtures. Crude oils contain many different hydrocarbon compounds that vary in appearance and
composition from one oil field to another. An "average" crude oil contains about 84% carbon;
14% hydrogen; and less than 2% sulfur, nitrogen, oxygen, metals, and salts (OSHA, 2003).
3-7
-------
In 2004, the petroleum refining industry used 5.6 billion barrels of crude oil in the
production of finished petroleum products (El A, 2005).'
Common Refinery Chemicals. In addition to crude oil. a variety of chemicals are used
in the production of petroleum products. The specific chemicals used will depend on specific
characteristics of the product in question. Table 3-1 lists the most common chemicals used by
petroleum refineries, their characteristics, and their applications.
In 2004, the petroleum refining industry used 581 million barrels of natural gas liquids
and other liquids in the production of finished petroleum products (E1A, 2005).
3.2.1.4 Types of Product Outputs
The petroleum refining industry produces a number of products that tend to fall into one
of three categories: fuels, finished nonfuel products, and feedstock for the petrochemical
industry. Table 3-2 briefly describes these product categories. A more detailed discussion of
petroleum fuel products can be found in Section 3.3.
Table 3-1. Types and Characteristics of Raw Materials used in Petroleum Refineries
Type Description
Crude Oil Heterogeneous mixture of different hydrocarbon compounds.
Oxygenates Substances which, when added to gasoline, increase the amount of oxygen in that
gasoline blend. Ethanol, methyl tertiary butyl ether (MTBE), ethyl tertiary butyl
ether (ETBE), and methanol are common oxygenates.
Caustics Caustics are added to desalting water to neutralize acids and reduce corrosion.
They are also added to desalted crude in order to reduce the amount of corrosive
chlorides in the tower overheads. They are used in some refinery treating processes
to remove contaminants from hydrocarbon streams. '
Leaded Gasoline Additives Tetraethyl lead (TEL) and tetramethyl lead (TML) are additives formerly used to
improve gasoline octane ratings but are no longer in common use except in
aviation gasoline
Sulfuric Acid and Sulfuric acid and hydrofluoric acid are used primarily as catalysts in alkylation
Hydrofluoric Acid processes. Sulfuric acid is also used in some treatment processes.
Source: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23, 2006.
1 A barrel is a unit of volume that is equal to 42 U.S. gallons.
3-8
-------
Table 3-2. Major Refinery Product Categories
Product Category - Description
Fuels Finished Petroleum products that are capable of releasing energy. These products
power equipment such as automobiles, jets, and ships. Typical petroleum fuel
products include gasoline, jet fuel, and residual fuel oil.
Finished nonfuel products Petroleum products that are not used for powering machines or equipment. These
products typically include asphalt, lubricants (such as motor oil and industrial
greases), and solvents (such as benzene, toluene, and xylene).
Feedstock Many products derived from crude oil refining, such as ethylene, propylene,
butylene, and isobutylene, are primarily intended for use as petrochemical
feedstock in the production of plastics, synthetic fibers, synthetic rubbers, and other
products.
Source: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23, 2006.
3.2.2 Emissions and Controls in Petroleum Refining
Petroleum refining leads to emissions of metals: spent acids; numerous toxic organic
compounds; and gaseous pollutants, including carbon monoxide (CO), sulfur oxides, (SOx),
nitrogen oxides (NOx), particulates, ammonia (NH3), hydrogen sulfide (H2S), and volatile
organic compounds (VOCs).
3.2.2.1 Gaseous and VOC Emissions
As previously mentioned, CO, SOx, NOx, NH3, and H2S emissions are produced along
with petroleum products. Sources of these emissions from refineries include fugitive emissions
of the volatile constituents in crude oil and its fractions, emissions from the burning of fuels in
process heaters, and emissions from the various refinery processes themselves.
Fugitive emissions occur as a result of leaks throughout the refinery. Although individual
leaks may be small, the sum of all leaks can result in a lot of hazardous emissions. These
emissions can be reduced by purchasing leak-resistant equipment and maintaining an ongoing
leak detection and repair program (EPA, 1995).
The numerous process heaters used in refineries to heat process streams or to generate
steam (boilers) for heating or other uses can be potential sources of SOx, NOx, CO, and
hydrocarbons emissions. Emissions are low when process heaters are operating properly and
using clean fuels such as refinery fuel gas, fuel oil, or natural gas. However, if combustion is not
complete, or the heaters are fueled using fuel pitch or residuals, emissions can be significant
(EPA, 1995).
3-9
-------
The majority of gas streams exiting each refinery process contain varying amounts of
refinery fuel gas, H2S, and NH3. These streams are directed to the gas treatment and sulfur
recovery units described in the previous section. Here, refinery fuel gas and sulfur are recovered
using a variety of processes. These processes create emissions of their own, which normally
contain H2S, SOx, and NOx gases (EPA, 1995).
Emissions can also be created by the periodic regeneration of catalysts that are used in
downstream processes. These processes generate streams that may contain relatively high levels
of CO, particulates, and VOCs. However, these emissions are treated before being discharged to
the atmosphere. First, the emissions are processed through a CO boiler to bum CO and any
VOCs, and then through an electrostatic precipitator or cyclone separator to remove particulates
(EPA, 1995).
3.2.2.2 Wastewater and Other Wastes
Petroleum refining operations produce a variety of wastewaters including process water
(water used in process operations like desalting), cooling water (water used for cooling that does
not come into direct contact with the oil), and surface water runoff (resulting from spills to the
surface or leaks in the equipment that have collected in drains). This wastewater typically
contains a variety of contaminants (such as hydrocarbons, suspended solids, phenols, NH3,
sulfides, and other compounds) and is treated in on-site facilities before being recycled back into
the production process or discharged.
Other wastes include forms of sludges, spent process catalysts, filter clay, and incinerator
ash. These wastes are controlled through a variety of methods including incineration, land filling,
and neutralization, among other treatment methods (EPA, 1995).
3.2.3 Costs of Production
Between 1995 and 2006, expenditures on input materials accounted for the largest cost to
petroleum refineries—amounting to 94% of total expenses (Figure 3-5). These material costs
included the cost of all raw materials, containers, scrap, and supplies used in production or repair
during the year, as well as the cost of all electricity and fuel consumed.
3-10
-------
Average Percentage
(1995-2006)
Materials
94%
Payroll
3%
Total Capital
3%
Figure 3-5. Petroleum Refinery Expenditures
Labor and capital accounted for the remaining expenses faced by petroleum refiners.
Capita! expenditures include permanent additions and alterations to facilities and machinery and
equipment used for expanding plant capacity or replacing existing machinery. A detailed
breakdown of how much petroleum refiners spent on each of these factors of production over
this l l-year period is provided in Table 3-3. A more exhaustive assessment of the costs of
materials used in petroleum refining is provided in Table 3-4.
3.3 The Demand Side
Estimating the economic impact the regulation will have on the petroleum refining
industry also requires characterizing various aspects of the demand for finished petroleum
products. This section describes the characteristics of finished petroleum products, their uses and
consumers, and possible substitutes.
3-ll
-------
Table 3-3. Labor, Material, and Capital Expenditures for Petroleum Refineries
(NAICS 324110)
Payroll (Smillions)
Materials (S
millions)
Total Capital (Smillions)
Year
Reported
2005
Reported
2005
Reported
2005
1995
3,791
4,603
112,532
136,633
5,937
7,209
1996
3,738
4,435
132,880
157,658
5,265
6,247
1997
3,885
4,595
127,555
150,865
4,244
5,020
1998
3,695
4,415
92,212
110,187
4,169
4,982
1999
3,983
4,682
114,131
134,146
3,943
4,635
2000
3,992
4,509
180,568
203,967
4,685
5,292
2001
4,233
4,743
158,733
177,838
6,817
7,638
2002
4,386
4,947
166,368
187,646
5,152
5,811
2003
4,752
5,227
185,369
203,893
6,828
7,510
2004
5,340
5,635
251,467
265,369
^ 6,601
6,966
2005
5,796
5,796
345,207
345,207
10,525
10,525
2006
5,984
5,751
396,980
381,546
11,175
10,741
Note: Adjusted for inflation using the producer price index industry for total manufacturing industries (Table 5-6).
Sources: U.S. Department of Commerce, Bureau of the Census. 2007. 2006 Annual Survey of Manufactures.
Obtained through American Fact Finder Database < http://factfinder.census.gov/home/saff/main.html?_lang=en>.
U.S. Department of Commerce, Bureau of the Census. 2006.2005 Annual Survey of Manufactures. M05(AS)-1.
Washington, DC: Government Printing Office. Available at
. As obtained on October 23, 2007.
U.S. Department of Commerce, Bureau of the Census. 2003a. 2001 Annual Survey of Manufactures. M01(AS)-1.
Washington, DC: Government Printing Office. Available at . As obtained on October 23, 2006.
U.S. Department of Commerce, Bureau of the Census. 2001.1999 Annua! Survey of Manufactures. M99(AS)-1
(RV). Washington, DC: Government Printing Office. Available at . As obtained on October 23,2006.
U.S. Department of Commerce, Bureau of the Census. 1998.1996 Annual Survey of Manufactures. M96(AS)-1
(RV). Washington, DC: Government Printing Office. Available at . As obtained on October 23,2006.
U.S. Department of Commerce, Bureau of the Census. 1997.1995 AnnuaI Survey of Manufactures. M95(AS)-1.
Washington, DC: Government Printing Office. Available at . As obtained on October 23,2006.
3.3.1 Product Characteristics
Petroleum refining firms produce a variety of different products. The characteristics these
products possess largely depend on their intended use. For example, the gasoline fueling our
automobiles has different characteristics than the oil lubricating the car's engine. However, as
discussed in Section 3.1.4, finished petroleum products can be categorized into three broad
groups based on their intended uses (EIA, 1999a):
¦ fuels—petroleum products that are capable of releasing energy such as motor
gasoline
3-12
-------
Table 3-4. Costs of Materials Used in Petroleum Refining Industry
2002
1997
Percentage
Percentage
Delivered
of Material
Delivered
of Material
Material
Cost (S106)
Costs
Cost ($106)
Costs
Petroleum Refineries NAICS 324110
Total materials
157,415,200
100.0%
118,682,535
100.0%
Domestic crude petroleum, including lease
63,157,497
40.1%
47,220,759
39.8%
condensate
Foreign crude petroleum, including lease
69,102,574
43.9%
48,172,988
40.6%
condensate
Foreign unfinished oils (received from
2;297,967
1.5%
2,373,376
2.0%
foreign countries for further processing)
Ethane (C2) (80% purity or more)
D
D
Propane (C3) (80% purity or more)
118,257
0.1%
269,928
0.2%
Butane (C4) (80% purity or more)
1,925,738
1.2%
1,567,875
1.3%
Gas mixtures (C2, C3, C4)
1,843,708
1.2%
952,009
0.8%
Isopentane and natural gasoline
810,530
0.5%
1,381,100
1.2% •
Other natural gas liquids, including plant
455,442
0.3%
1,427,123
1.2%
condensate
Toluene and xylene (100% basis)
159,563
0.1%
N
Additives (including antioxidants,
40,842
0.0%
262,228
0.2%
antiknock compounds, and inhibitors)
Other additives (including soaps and
709
0.0%
200,005
0.2%
detergents)
Animal and vegetable oils
D
D
Chemical catalytic preparations
D
647,040
0.5%
Sodium hydroxide (caustic soda) (100%
129,324
0.1%
41,741
0.0%
NaOH)
Sulfuric acid, excluding spent (100%
189,912
o.i%
56,514
0.0%
H2S04)
Metal containers
9,450
0.0%
60,531
0.1%
Plastics containers
D
N
Paper and paperboard containers
D
18,404
0.0%
Cost of materials received from petroleum
8,980,758
5.7%
4,981,370
4.2%
refineries and lube manufacturers
All other materials and components, parts,
5,722,580
3.6%
4,233,383
3.6%
containers, and supplies
Materials, ingredients, containers, and
576,175
0.4%
4,779,890
' 4.0%
supplies, nsk
Source: U.S. Department of Commerce, Bureau of the Census. 2004. 2002 Economic Census, Industry Series—
Shipbuilding and Repair. Washington, DC: Government Printing Office. Available at
-------
A list of selected products from each of these groups is presented in Table 3-5 along with a
description of each product's characteristics and primary uses.
Table 3-5. Major Refinery Products
Product
Description
Fuels
Gasoline
Kerosene
Liquefied petroleum gas
(LPG)
Distillate fuel oil
Residual fuels
Petroleum coke
A blend of refined hydrocarbons, motor gasoline ranks first in usage among petroleum
products. It is primarily used to fuel automobiles and lightweight trucks as well as
boats, recreational vehicles, lawn mowers, and other equipment. Other forms of
gasoline include Aviation gasoline, which is used to power small planes.
Kerosene is a refined middle-distillate petroleum product that finds considerable use
as a jet fuel. Kerosene is also used in water heaters, as a cooking fuel, and in lamps.
LPG consists principally of propane (CjHg) and butane (C4H10). It is primarily used
as a fuel in domestic heating, cooking, and farming operations.
Distillate fuel oil includes diesel oil, heating oils, and industrial oils. It is used to
power diesel engines in buses, trucks, trains, automobiles, as well as other machinery.
Residual fuels are the fuels distilled from the heavier oils that remain after
atmospheric distillation, they find their primary use generating electricity in electric
utilities. However, residual fuels can also be used as fuel for ships, industrial boiler
fuel, and commercial heating fuel.
Coke is a high carbon residue that is the final product of thermal decomposition in the
condensation process in cracking. Coke can be used as a low-ash solid fuel for power
plants.
Finished Nonfuel Products
Coke In addition to use as a fuel, petroleum coke can be used a raw material for many
carbon and graphite products such as fumace electrodes and liners.
Asphalt Asphalt, used for roads and roofing materials, must be inert to most chemicals and
weather conditions.
Lubricants Lubricants are the result of a special refining process that produce lubricating oil base
stocks, which are mixed with various additives. Petroleum lubricating products
include spindle oil, cylinder oil, motor oil, and industrial greases.
Solvents A solvent is a fluid that dissolves a solid, liquid, or gas into a solution. Petroleum
based solvents, such as Benzyme, are used top manufacture detergent and synthetic
fibers. Other solvents include toluene and xylene.
Feedstock
Ethylene Ethylene is the simplest alkene and has the chemical formula C2H4. It is the most
produced organic compound in the world and it is used in the production of many
products. For example, one of ethylene's derivatives is ethylene oxide, which is a
primary raw material in the production of detergents.
Propylene Propylene is an organic compound with the chemical formula C3H6. It is primarily
used the production of polypropylene, which is used in the production of food
packaging, ropes, and textiles.
Sources: U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003. OSHA
Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-00-015. Washington, DC: U.S.
DOL. Available at . As obtained on October 23, 2006.
U.S. Department of Energy, Energy Information Administration (EIA). 1999.
3-14
-------
3.3.2 Uses and Consumers
Finished petroleum products are rarely consumed as final goods in themselves. Instead,
they are used as primary inputs in the creation of a vast number of other goods and services. For
example, goods created from petroleum products include fertilizers, pesticides, paints, thinners,
cleaning fluids, refrigerants, and synthetic fibers (EPA, 1995). Similarly, fuels made from
petroleum are used to run vehicles and industrial machinery and generate heat and electrical
power. As a result, the demand for many finished petroleum products is derived from the .
demand for the goods and services they are used to create.
The principal end users of petroleum products can be separated into five sectors:
¦ Residential sector—private homes and residences
¦ Industrial sector—manufacturing, construction, mining, agricultural, and forestry
establishments
• Transportation sector—private and public vehicles that move people and
commodities such as automobiles, ships, and aircraft
¦ Commercial sector—nonmanufacturing or nontransportation business
establishments such as hotels, restaurants, retail stores, religious and nonprofit
organizations, as well federal, state, and local government institutions
¦ Electric utility sector—privately and publicly owned establishments that generate,
transmit, distribute, or sell electricity (primarily) to the public; nonutility power
producers are not included in this sector
Of these end users, the transportation sector consumes the largest share of petroleum
products, accounting for 67% of total consumption in 2005 (El A, 2006a). In fact, petroleum
products like motor gasoline, distillate fuel, and jet fuel provide virtually all of the energy
consumed in the transportation sector (El A, 1999a).
Of the three petroleum product categories, end-users primarily consume fuel. Fuel
products account for 9 out of 10 barrels of petroleum used in the United States (EIA, 1999a). In
2005, motor gasoline alone accounted for 49% of demand for finished petroleum products (EIA,
2006a).
3.3.3 Substitution Possibilities in Consumption
A major influence on the demand for finished petroleum products is the availability of
substitutes. In some sectors, like the transportation sector, it is currently difficult to switch
quickly from one fuel to another without costly and irreversible equipment changes, but other
sectors can switch relatively quickly and easily (EIA, 1999a).
3-15
-------
For example, equipment at large manufacturing plants often can use either residual fuel
oil or natural gas. Often coal and natural gas can be easily substituted for residual fuel oil at
electricity utilities. As a result, we would expect demand in these industries to be more sensitive
to price (in the short run) than in others (EIA, 1999a).
However, over time, demand for petroleum products could become more elastic. For
example, automobile users could purchase more fuel-efficient vehicles or relocate to areas that
would allow them to make fewer trips. Technological advances could also create new products
that compete with petroleum products that currently have no substitutes. An example of such a
technological advance would be the invention of ethanol (an alcohol produced from biomass).
which can substitute for gasoline in spark-ignition motor vehicles (EIA, 1999a).
3.4 Industry Organization
This section examines the organization of the U.S. petroleum refining industry, including
market structure, firm characteristics, plant location, and capacity utilization. Understanding the
industry's organization helps determine how it will be affected by new emissions standards.
3.4.1 Market Structure
Market structure characterizes the level and type of competition among petroleum
refining companies and determines their power to influence market prices for their products. For
example, if an industry is perfectly competitive, then individual producers cannot raise their
prices above the marginal cost of production without losing market share to their competitors.
Understanding pricing behavior in the petroleum refining industry is crucial for performing
subsequent EIAs.
According to basic microeconomic theory, perfectly competitive industries are
characterized by unrestricted entry and exit of firms, large numbers of firms, and undifferentiated
(homogenous) products being sold. Conversely, imperfectly competitive industries or markets t
are characterized by barriers to entry and exit, a smaller number of firms, and differentiated
products (resulting from either differences in product attributes or brand name recognition of
products). This section considers whether the petroleum refining industry is competitive based
on these three factors.
3.4.1.1 Barriers to Entry
Firms wanting to enter the petroleum refining industry may face at least two major
barriers to entry. First, according to a 2004 Federal Trade Commission staff study, there are
significant economies of scale in petroleum refinery operations. This means that costs per unit
3-16
-------
fall as a refinery produces more finished petroleum products. As a result, new firms that must
produce at relatively low levels will face higher average costs than firms that are established and
produce at higher levels, which will make it more difficult for these new firms to compete
(Nicholson. 2005). This is known as a technical barrier to entry.
Second, legal barriers could also make it difficult for new firms to enter the petroleum
.refining industry. The most common example of a legal barrier to entry is patents—intellectual
property rights, granted by the government, that give exclusive monopoly to an inventor over his
invention for a limited time period. In the petroleum refining industry, firms rely heavily on
process patents to appropriate returns from their innovations. As a result, firms seeking to enter
the petroleum refining industry must develop processes that respect the novelty requirements of
these patents, which could potentially make entry more difficult for new firms (Langinier, 2004).
A second example of a legal barrier would be environmental regulations that apply only to new
entrants or new pollution sources. Such regulations would raise the operating costs of new firms
without affecting the operating costs of existing ones. As a result, new firms may be less
competitive.
Although neither of these barriers are impossible for new entrants to overcome, they can
make it more difficult for new firms to enter the market for manufactured petroleum products. As
a result, existing petroleum refiners could potentially raise their prices above competitive levels
with less worry about new firms entering the market to compete away their customers with lower
prices. It was not possible during this analysis to quantify how significant these barriers would be
for new entrants or what effect they would have on market prices. However, existing firms
would still face competition from each other. In an unconcentrated industry, competition among
existing firms would work to keep prices at competitive levels.
3.4.1.2 Measures of Industry Concentration
Economists often use a variety of measures to assess the-concentration of a given
industry. Common measures include four-firm concentration ratios (CR4), eight-firm
concentration ratios (CR8), and Herfindahl-Hirschmann indexes (HH1). The CR4s and CR8s
measure the percentage of sales accounted for by the top four and eight firms in the industry. The
HHls are the sums of the squared market shares of firms in the industry. These measures of
industry concentrated are reported for the petroleum refining industry (NAICS 324110) in
Table 3-6 for selected years between 1985 and 2003.
3-17
-------
Table 3-6. Market Concentration Measures of the Petroleum Refining Industry: 1985 to
2003
Measure
1985
1990
1996
2000
2001
2002
2003
Herfindahl-Hirschmann Index (HHI)
493
437
412
611
686
743
728
Four-firm concentration ratio (CR4)
34.4
31.4
27.3
40.2
42.5
45.4
44.4
Eight-firm concentration ratio (CR8)
54.6
52.2
48.4
61.6
67.2
70.0
69.4 -
Source: Federal Trade Commission (FTC). 2004. "The Petroleum Industry: Mergers, Structural Change, and
Antitrust Enforcement." Available ait . As obtained on
February 6, 2007.
Between 1990 and 2000, the HHI rose from 437 to 611, which indicates an increase in
market concentration over time. This increase is partially due to merger activity during this time
period. Between 1990 and 2000, over 2,600 mergers occurred across the petroleum industry;
13% of these mergers occurred in the industry's refining and marketing segments (GAO, 2007).
Unfortunately, there is no objective criterion for determining market structure based on
the values of these concentration ratios. However, accepted criteria have been established for
determining market structure based on the HHIs for use in horizontal merger analyses (U.S.
Department of Justice and the Federal Trade Commission, 1992). According to these criteria,
industries with HHIs below 1,000 are considered unconcentrated (i.e., more competitive);
industries with HHIs between 1,000 and 1,800 are considered moderately concentrated (i.e.,
moderately competitive); and industries with higher HHIs are considered heavily concentrated.
Based on this criterion, the petroleum refining industry continues to be unconcentrated even after
an increase in merger activity.
A more rigorous examination of market concentration was conducted in a 2004 Federal
Trade Commission (FTC) staff study. This study explicitly accounted for the fact that a refinery
in one geographic region may not exert competitive pressure on a refinery in another region if
transportation costs are high. This was done by comparing HHIs across Petroleum
Administration for Defense Districts (PADDs). PADDs separate the United States into five
geographic regions or districts. They were initially created during World War II to help manage
the allocation of fuels during wartime. However, they have remained in use as a convenient way
of organizing petroleum market information (FTC, 2004).
This study concluded that these geographic markets were not highly concentrated.
PADDs 1, II, and 111 (East Coast, Midwest, and Gulf Coast) were sufficiently connected that they
exerted a competitive influence on each other. The HHI for these combined regions was 789 in
2003, indicating a low concentration level. Concentration in PADD IV (Rocky Mountains) was
3-18
-------
also low in 2003, with an HHI of 944. PADD V gradually grew more concentrated in the 1990s
after a series of significant refinery mergers. By 2003, the region's HHI was 1,246, indicating
growth to a moderate level of concentration (FTC, 2004).
3.4.1.3 Product Differentiation
Another way firms can influence market prices for their product is through product
differentiation. By differentiating one's product and using marketing to establish brand loyalty,
manufacturers can raise their prices above marginal cost without losing market share to their
competitors.
While we saw in Section 3.3 that there are a wide variety of petroleum products with
many different uses, individual petroleum products are by nature quite homogenous. For
example, there is little difference between premium motor gasoline produced at different
refineries (Mathtech, 1997). As a result, the role of product differentiation is probably quite
small for many finished petroleum products. However, there are examples of relatively small
refining businesses producing specialty products for small niche markets. As a result, there may
be some instances where product differentiation is important for price determination.
3.4.1.4 Competition among Firms in the Petroleum Refining Industry
Overall, the petroleum industry is characterized as producing largely generic products for
sale in relatively unconcentrated markets. Although it is not possible to quantify how much
barriers to entry and other factors will affect competition among firms, it seems unlikely that
individual petroleum refiners would be able to significantly influence market prices given the
current structure of the market.
3.4.2 Characteristics of U.S. Petroleum Refineries and Petroleum Refining Companies
A petroleum refinery is a facility where labor and capital are used to convert material
inputs (such as crude oil and other materials) into finished petroleum products. Companies that
own these facilities are legal business entities that conduct transactions and make decisions that
affect the facility. The terms "facility," "establishment," and "refinery" are synonymous in this
\
study and refer to the physical location where products are manufactured. Likewise, the terms
"company" and "firm" are used interchangeably to refer to the legal business entity that owns
one or more facilities. This section presents information on refineries, such as their location and
capacity utilization, as well as financial data for the companies that own these refineries.
3-19
-------
3.4.2.1 Geographic Distribution of U.S. Petroleum Refineries
There are approximately 149 petroleum refineries operating in the United States, spread
across 33 states. The number of petroleum refineries located in each of these states is listed in
Table 3-7. This table illustrates that a significant portion of petroleum refineries are located
along the Gulf of Mexico region. The leading petroleum refining states are Texas, California,
and Louisiana.
Table 3-7. Number of Petroleum Refineries, by State
State Number of Petroleum Refineries
Alabama
4
Alaska
6
Arkansas
2
California
21
Colorado
2
Delaware
1
Georgia
1
Hawaii
2
Illinois
4
Indiana
2
Kansas
3
Kentucky
2
Louisiana
18
Michigan
1
Minnesota
2
Mississippi
4.
Montana
4
Nevada
1
New Jersey
6
New Mexico
3
North Dakota
1
Ohio
4
Oklahoma
5
Oregon
1
Pennsylvania
5
Tennessee
1
Texas
25
Utah
5
Virginia
1
Washington
5
West Virginia
1
Wisconsin .
1
Wyoming
5
Total
149
Source: U.S. Department of Energy, Energy Information Administration (El A). 2006b. "Refinery Capacity Report
2006." Available at . As obtained on October 23, 2006.
3-20
-------
3.4.2.2 Capacity Utilization
Capacity utilization indicates how well current refineries meet demand. One measure of
capacity utilization is capacity utilization rates. A capacity utilization rate is the ratio of actual
production volumes to full-capacity production volumes. For example, if an industry is
producing as much output as possible without adding new floor space for equipment, the
capacity utilization rate would be 100 percent. On the other hand, if under the same constraints
the industry were only producing 75 percent of its maximum possible output, the capacity
utilization rate would be 75 percent. On an industry-basis, capacity utilization is highly variable
from year to year depending on economic conditions. It is also variable on a company-by-
company basis depending not only on economic conditions, but also on company's strategic
position in its particular industry. While some plants may have idle production lines or empty
floor space, others need additional space or capacity.
Table 3-8 lists the capacity utilization rates for petroleum refineries from 2000 to 2006. It
is interesting to note the significant drop in capacity utilization in 2005. This would seem counter
intuitive since there does not appear to be evidence that demand for petroleum products is not
dropping. To understand why this might be the case, one must first realize that the capacity
utilization ratio in petroleum industry represents the utilization of the atmospheric crude oil
distillation units.
Table 3-8. Full Production Capacity Utilization Rates for Petroleum Refineries
Year
Petroleum Refineries
Capacity Utilization Rates
(NA1CS 324110)
Gross Input to Atmospheric
Crude Oil Distillation Units
(1,000s of barrels per day)
Operational Capacity
(1,000s of barrels per day)
2000
92.6
15,299
16,525
2001
92.6
15,352
16,582
2002
90.7
15,180
16,744
2003
92.6
15,508
16,748
2004
93.0
15,783
16,974
2005
90.6
15,578
17,196
2006
89.7
15,602
17,385
Source: U.S. Department of Energy, Energy Information Administration (EIA). 2007a. "Refinery Utilization and
Capacity." Available at . As obtained on
January, 2007.
This is calculated for the petroleum industry by dividing the gross input to atmospheric
crude oil distillation units (all inputs involved in atmospheric crude oil distillation, such as crude
oil) by the industry's operational capacity.
3-21
-------
In 2004. operational capacity increased from 16,974,000 barrels per calendar day to
17,196,000 barrels per calendar day. However, gross inputs fell from 15,783,000 barrels per
calendar day in 2004 to 15,578.000 in 2005. This indicates that capacity utilization sagged due to
a drop in production inputs. In 2006, gross inputs grew 0.15% to 15,602,000 barrels per day.
However, since operational capacity grew much faster (from 17,196,000 to 17,385,000 or
1.00%), capacity utilization rates for the industry continued to fall.
3.4.2.3 Characteristics of Small Businesses Owning U.S. Petroleum Refineries
According to the Small Business Administration (SBA), a small business in the
petroleum refining industry is defined for government procurement purposes as having 1,500 or
fewer employees and an Operable Atmospheric Crude Oil Distillation capacity of no more than
125,000 barrels per calendar day total (SBA, 2008). We applied this definition in defining a
small business refiner in our proposal R1A. However, as part of a response to a comment made
on the proposal, we define a small business as only as having 1,500 or fewer employees.
As of January 2006, there were 149 petroleum refineries operating in the continental
United States with a cumulative capacity of processing over 17 million barrels of crude per
calendar day (EIA, 2006c). RT1 identified 58 parent companies owning refineries in the United
States and was able to collect employment and sales data for 47 (84%) of them.
The distribution of employment across companies is illustrated in Figure 3-6. As this
figure shows, 25 companies (53%) of these 47 employ fewer than 1,500 workers and would be
considered small businesses. These firms earned an average of $1.04 billion of revenue per year,
while firms employing more than 1,500 employees earned an average of $84.2 billion of revenue
per year (Figure 3-7). A distribution of the number of firms earning different levels of revenue is
presented in Figure 3-8.
Employment, crude capacity, and location information are provided in Table 3-9 for each
of companies employing 1,500 employees or less. Similar information can be found for all 56
companies owning petroleum refineries in Appendix A.
In Section 3.4.2.1, we discussed how petroleum refining operations are characterized by
economies of scale—that the cost per unit falls as a refinery produces more finished petroleum
products. This means that smaller petroleum refiners face higher per unit costs than larger
refining operations because they produce fewer petroleum products. As a result, some smaller
firms have sought to overcome their competitive disadvantage by locating close to product-
consuming areas to lower transportation costs and serving niche product markets (FTC, 2004).
3-22
-------
60%
50%
g 40%
il 30%
H-
o 20%
| 10%
0%
<100 100-250 250-500 500-1000 1000-
1500
>1500
Figure 3-6. Employment Distribution of Companies Owning Petroleum Refineries (N=47)
Sources: Dun & Bradstreet. 2007a. 2007 D&B Million Dollar Directory. Pennsylvania: Dun & Bradstreet Inc.
Dun & Bradstreet Small Business Solutions. Small Business Database. Available at
.
Gale Research Inc. 2007. Ward's Business Directory of U.S. Private and Public Companies. Detroit: Gale
Research.
Hoovers. 2007. Free Content, Company Information. Available at .
| $100,000
E $80,000
C
03
a:
03
O)
as
Im.
03
>
<
$60,000
$40,000
$20,000
$0
<100 100-250 250-500 500-1000 1000- >1500
1500
Figure 3-7. Average Revenue of Companies Owning Petroleum Refineries by Employment
(N=47)
Sources:Dun & Bradstreet. 2007. 2007 D&B Million Dollar Directory. Pennsylvania: Dun & Bradstreet Inc.
Dun & Bradstreet Small Business Solutions. Small Business Database. Available at
.
Gale Research Inc. 2007. Ward's Business Directory of U S Private and Public Companies. Detroit: Gale
Research.
Hoovers. 2007. Free Content, Company Information. Available at .
3-23
-------
70%
^ 60%
0s-
w 50%
E
£ 40%
o 30%
£ 20%
W 10%
0%
Figure 3-8. Revenue Distribution of Companies Owning Petroleum Refineries (N=47)
Sources: Dun & Bradstreet, 2007. 2007 D&B Million Dollar Directory. Pennsylvania: Dun & Bradstreet Inc.
Dun & Bradstreet Small Business Solutions. Small Business Database. Available at
.
Gale Research Inc. 2007. Ward's Business Directory of U S Private and Public Companies. Detroit: Gale
Research.
Hoovers. 2007. Free Content, Company Information. Available at .
A good example of a firm locating close to prospective customers is Countrymark
Cooperative, Inc., which was started in the 1930s for the express purpose of providing farmers in
Indiana with a consistent supply of fuels, lubricants, and other products. A good example of a
firm producing niche products is Calumet Lubricants, which focuses on developing and
manufacturing naphthenic specialty oils.
However, recent developments are making these factors less important for success in the
industry. For example, the entry of new product pipelines is eroding the locational advantage of
smaller refineries (FTC, 2004). This trend can possibly be illustrated by the fact that most
refineries owned by small businesses tend to be located in relatively rural areas (see Table 3-9).
The median population density of counties occupied by small refineries is 94 people per square
mile. This could suggest that refineries do not rely on the population surrounding them to support
their refining operations.
To obtain a better sense of where the customers of these small refiners are located, RT1
spoke with representatives from four different companies. Three of these representatives
indicated that they primarily serve customers outside their local areas. In particular, two of these
businesses were primarily fuel producers that used pipelines to deliver their product to customers
up to 300 miles away.
.— I if I
, ,
<$5 $5-$10 $10-$50 $50- $100- $500- >$1000
$100 $500 $1000
3-24
-------
Table 3-9. Characteristics of Small Businesses in the Petroleum Refining Industry
Parent Company
Parent
Company
Type
Cummula-
tive Crude
Capacity
(bbl/cd)
Parent
Company
Sales
(SMillions)
Parent
Company
Employ-
ment (#)
Facility Name
Facility City
Facility
State
Facility
County
County ID
Facility
County
Population
Density
(2000)
AGE Refining &
Private
12,200
287
52
AGE Refining &
San Antonio
TX
Bexar County
TXBexar
1,117
Manufacturing
Manufacturing
County
American Refining
Private
10,000
350
310
American Refining
Bradford
PA
McKean
PAMcKean
47
Group
Group
County
County
Arabian American
Public
0
80
118
South Hampton
Silsbee
TX
Hardin
TXHardin
54
Development Co
Resources Inc.
County
County
T
{Calcasieu Refining
Private
30,000
638
" 51
Calcasieu Refining
Lake Charles
LA
Calcasieu
LACalcasieu
171
jCo.
Co.
Parish
Parish
Calumet Specialty
Public
63,320
1,641
, 350
Calumet Specialty
Shreveport
LA
Caddo Parish
LACaddo Parish
286
i Products
Products
t
Calumet Specialty
Cotton Valley
LA
Caddo Parish
LACaddo Parish
286
Products
Calumet Specialty
Princeton
LA
Caddo Parish
LACaddo Parish
286
}
Products
jCountrymark
Private
23,000
87
300
Countrymark
Mt. Vernon
IN
Posey County
INPosey County
66
'Cooperative, Inc.
Cooperative, Inc.
| Cross Oil & Refining
Private
7,200
49
110
Cross Oil &
Sniackover
AR
Union County
ARUnion
44
[Co. Inc.
Refining Co. Inc.
County
CVR Energy Inc.
Public
112,000
3,038
577
Coffeyville
Coffeyville
KS
Montgomery
KSMontgomery
56 "
Resources LLC
County
County
Foreland Refining
Private
2.000
56 "
100
Foreland Refining
Tonopah/Eagle
NV
Nye County
NVNye County
2
Co.
Co.
Springs
Frontier Oil Corp
Private
153,000
4,000
727
Frontier Oil &
Cheyenne
WY
Laramie
WYLaramie
30 '
Refining Co.
County
County
Frontier Oil Corp
El Dorado
KS
Butler County
KSButler
42
County
Gary-Williams Co
Private
54,000
"97
200
Wynnewood
Wynnewood
OK
Garvin
OKGarvin
34
Refining Co.
County
County
(Continued)
-------
Table 3-9. Characteristics of Small Businesses in the Petroleum Refining Industry (continued)
Facility
Cumulative
Parent
Parent
County
Parent
Crude
Company
Company
Population
Company
Capacity
Sales
Employ-
Facility
Facility
Density
Parent Company
Type
(bbl/cd)
(SMillions)
ment (#)
Facility Name
Facility City
State
County
(2000)
Goodway Refining LLC
Private
4.100
3
18
Goodway Refining
Atmore
AL
Escambia
41
LLC
County
Greka Integrated Inc
Private
9.500
22
145
Greka Integrated Inc
Santa Maria
CA
Santa Barbara
146
County
Gulf Atlantic
Private
16,700
9
32
Gulf Atlantic
Mobile Bay
AL
Mobile County
324
Operations LLC
Operations LLC
Holly Corp.
Public
99,700
4,023
859
Holly Corp.
Woods Cross
UT
Davis County
785
Navajo Refining Co.
Artesia
NM
Eddy County
12
Hunt Refining Co.
Private
45,500
" 4,871
1,100
Hunt Refining Co.
Tuscaloosa
AL
Tuscaloosa
125
.
County
Hunt Southland
Lumberton
MS
Lamar County
79
Refining
Hunt Southland
Sandersville
MS
Lamar County
79
Refining
Lion Oil Co.
Private
70,000
247
425
Lion Oil Co.
El Dorado
AR
Union County
44
Pelican Refining Co.
Private
0
29
62
Pelican Refining Co.
Lake Charles
LA
Calcasieu
171
LLC
LLC '
Parish
Placid Refining Inc.
Private
56,000
1,400
200
Placid Refining Inc.
Port Allen
LA
West Baton
113
Rouge Parish
San Joaquin Refining
Private
15,000
288
20
San Joaquin Refining
Bakersfield
CA
Kern County
81
Co.. Inc.
Co., Inc.
Somerset Oil Inc
Private
5,500
55
150
Somerset Refinery
Somerset
KY
Pulaski County
85
Trigeant Ltd.
Private
0
5
50
Trigeant Ltd.
Corpus Christi
TX
Nueces County
375
(Continued)
-------
Table 3-9. Characteristics of Small Businesses in the Petroleum Refining Industry (continued)
Parent Company
Parent
Company
Type
Cumulative
Crude
Capacity
(bbl/cd)
Parent
Company
Sales
(SMillions)
Parent
Company
Employ-
ment (#)
Facility Name
Facility City
Facility
State
Facility
County
Facility
County
Population
Density
(2000)
Western Refining, Inc.
Public
212,200
4,200
416
Western Refining, Inc.
Giant Refining Co.
Giant Refining Co.
Giant Refining Co.
El Paso
Yorktown
Bloomfield
Gallup
TX
VA
NM
NM
El Paso County
Y ork County
San Juan
County
McKinley
County
671
533
21
14
World Oil Corp
Private
8,500
277
475
Lunday-Thagard Co.
South Gate
CA
Los Angeles
County
2,344
Wyoming Refining Co.
Private
12,500
340
107
Wyoming Refining
Co.
Newcastle
WY
Weston County
3
Total
2,128,860
59,738
12,688
Sources: Dun & Bradstreet. 2007. 2007 D&B Million Dollar Directory. Pennsylvania: Dun & Bradstreet Inc.
Dun & Bradstreet Small Business Solutions. Small Business Database. Available at .
Gale Research Inc. 2007. Ward's Business Directory of U S Private and Public Companies. Detroit: Gale Research.
Hoovers. 2007. Free Content, Company Information. Available at . As obtained on April 11, 2007.
U.S. Department of Commerce, Bureau of the Census. 2000. "Population Density by County: Census 2000 Summary File 1 (SF 1) 100-Percent Data".
Available through American Fact Finder < http://factfinder.census.gov/home/saff/main.html?_lang=en>. As obtained on February 21, 2008.
-------
Capacity information for the 29 refineries owned by small businesses also suggests that
fewer small businesses are focusing on developing specialty products or serving local customers
as major parts of their business plan. For example, in 2006 these 29 refineries had a collective
crude refining capacity of 778.920 barrels per calendar day or 857,155 barrels per stream day
(EI A, 2006c). Approximately 21% of this total capacity was devoted to producing specialty
products or more locally focused products such as aromatics, asphalt, lubricants, and petroleum
coke. The remaining 79% was used to produce gasoline, kerosene, diesel fuel, and liquefied
petroleum gases. As discussed in Section 3.4.1.3, fuel products tend to be quite homogenous
(gasoline from one refinery is not very different from gasoline from another refinery), and they
are also normally transported by pipeline.
3.5 Markets
This section provides data on the volume of petroleum products produced and consumed
in the United States, the quantity of products imported and exported, and the average prices of
major petroleum products. The section concludes with a discussion of future trends for the
petroleum refining industry.
3.5.1 U.S. Petroleum Consumption
Figure 3-9 illustrates the amount of petroleum products supplied between 2000 and 2006
(measured in millions of barrels of oil). These data represent the approximate consumption of
petroleum products because it measures the disappearance of these products from primary
sources (i.e., refineries, natural gas processing plants, blending plants, pipelines, and bulk
terminals).
Between 2000 and 2004, U.S. consumption of petroleum products increased by 5%.
Consumption grew steadily from 2001 and 2004 before leveling off and slightly declining in
2006 (Figure 3-9). This reduced growth was primarily the result of less jet fuel and residual fuel
being consumed in recent years (Table 3-10).
3.5.2 U.S. Petroleum Production
Table 3-11 reports the number of barrels of major petroleum products produced in the
United States between 2000 and 2006. U.S. production of petroleum products at refineries and
blenders grew steadily between 1995 and 2003. However, production declined by 0.35% in
2005. This drop was possibly the result of damage inflicted by two hurricanes (Hurricane Katrina
and Hurricane Rita) on the U.S. Gulf Coast—the location of many U.S. petroleum refineries
(Section 3.4.2). According to the American Petroleum Institute, approximately 30% of the U.S.
refining industry was shut down as a result of the damage (API, 2006). In 2006, production of
3-28
-------
Figure 3-9. Total Petroleum Products Supplied (millions of barrels per year)
Table 3-10. Total Petroleum Products Supplied (millions of barrels per year)
Year
Motor
Gasoline
Jet Fuel
Distillate
Fuel Oil
Residual
, Fuel Oil
Liquefied
Petroleum
Gases
Other
Products
Total
2000
3,101
631
1,362
333
816
967
7,211
2001
3.143
604
1,404
296
746
- 978
7,172
2002
3,229
591
1,378
255
789
969
7,213
2003
3,261
576
1,433
282
¦ 757
1,003
7,312
2004
3,333
597
1,485
316
780
1,076
7,588
2005
3,343
613
1,503
336
741
1,057
7,593
2006
3,377
596
1,522
251
749
1,055
7,551
Source: U.S. Department of Energy, Energy Information Administration (EIA). "Petroleum Supply Annuals 1996—
2007, Volume 1." Available at . As obtained on October 31,2007.
petroleum products rebounded, increasing 1% over 2004 levels. Additional production data are
presented in Table 3-12, which reports the value of shipments of products produced by the
petroleum refining industry between 1997 and 2006.
3.5.3 International Trade
International trade is a growing component of the U.S. Petroleum refining industry. This
trend is demonstrated in Tables 3-13 and 3-14. Between 1995 and 2006, imports and exports of
petroleum products increased by more than 50%. While imports of most major petroleum
3-29
-------
Table 3-11. U.S. Refinery and Blender Net Production (millions of barrels per year)
Year
Motor
Gasoline
Jet Fuel
Distillate
Fuel Oil
Residual
Fuel Oil
Liquefied
Petroleum
Gases
Other
Products
Total
2000
2,910
588
1,310
255
258
990
6,311
2001
2,928
558
1,349
263
243
968
6.309
2002
2,987
553
1,311
219
245
990
6,305
2003
2,991
543
1,353
241
240
1,014
6.383
2004
'3,025
566
1,396
240
• 236
1,057
6.520
2005
3,036
564
1,443
229
209
1,015
6.497
2006
3,053
541
1,475
232
229
1,032
6,561
Source: U.S. Department of Energy, Energy Information Administration (E1A). "Petroleum Supply Annuals 1996—
2007, Volume I." Available at .
U.S. Department of Commerce, Bureau of the Census. 2003b. 2001 Annual Survey of Manufactures. M01(AS)-2.
Washington, DC: Government Printing Office. Available at < http://www.census.gov/prod/2003pubs/m0Ias-
2.pdf. As obtained on March 4, 2008.
(
products grew at approximately the same rate, the growth of petroleum product exports was
driven largely by residual fuel oil and other petroleum products.
However, the United States remains a net importer of petroleum products. In 2006, the
United States imported nearly three times more petroleum products than it exported. These
imported petroleum products accounted for 17% of total petroleum products consumed that year
(1,310 millions of barrels per year/7,551 millions of barrels per year).
3-30
-------
Table 3-13. Imports of Major Petroleum Products (millions of barrels per year)
Year
Motor
Gasoline
Jet Fuel
Distillate
Fuel Oil
Residual
Fuel Oil
Liquefied
Petroleum
Gases
Other
Products
Total
1995
97
35
71
68
53
262
586
1996
123
40
84
91
61
322
721
1997
113
33
83
71
62
345
707
1998
114
45
77
101
71
324
731
1999
139
47
91
86
66
344
774
2000
156
59
108
129
79
343
874
2001
166
54
126
108
75
400
928
2002
182
39
98
91 .
67
396
872
2003
189
40
122
119
82
397
949
2004
182
47
119
156 .
96
520
1,119
2005
220
69
120
193
120
587
1,310
2006
173
68
133
128
121
687
1,310
Source: U.S. Department of Energy, Energy Information Administration (ElA). "Petroleum Supply Annuals 1996-
2007, Volume IAvailable at . As obtained on October 31,2007.
Table 3-14. Exports of Major Petroleum Products (millions of barrels per year)
Year
Motor
Gasoline
Jet Fuel
Distillate
Fuel Oil
Residual
Fuel Oil
Liquefied
Petroleum
Gases
Other
Products
Total
1995
38
8
67
49
21
128
312
1996
38
17
70
37
19
138
319
1997
50
13
56
44
18
147
327
1998
46
9
45
50
15
139
305
1999
40
11
59
47
18
124
300
2000
53
12
63
51
27
157
362
2001
48
10
44
70
16
159
347
2002
45
3
41
65
24
177
356
2003
46
7
39
72
20
186
370
2004
45
15
40
75
16
183
374
2005
49
19
51
92
19
183
414
2006
52
15
79
103
21
203
472
Source: U.S. Department of Energy, Energy Information Administration (EIA). "Petroleum Supply Annuals 1996—
2007, Volume 1." Available at . As obtained on October, 31, 2007.
3.5.4 Market Prices
The average nominal prices of major petroleum products sold to end. users are provided
for selected years in Table 3-15.' As these data illustrate, nominal prices rose substantially
between 2004 and 2006. In particular, the price of motor gasoline rose 48% over this 2-year
period.
1 Sales to end users are those made directly to the consumer of the product. This includes bulk consumers, such as
agriculture, industry, and utilities, as well as residential and commercial consumers.
3-31
-------
Table 3-15. Average Price of Major Petroleum Products Sold to End Users (cents per
gallon)
Product
1995
2000
2002
2004
2005
2006
Motor gasoline
76.5
110.6
94.7
143.5
182.9
212.8
No. 1 distillate fuel
' 62
98.8
82.8
126.2
183.2
213.7
No. 2 distillate fuel
56
93.4
75.9
123.5
177.7
209.1
Jet fuel
54
89.9
72.1
120.7
173.5
199.8
Residual fuel oil
39.2
60.2
56.9
73.9
104.8
121.8
Source: U.S. Department of Energy, Energy Information Administration (EI A). 2007b. "Refiner Petroleum Product
Prices by Sales Type." Available at . As
obtained on January 11, 2008.
Note: Prices do not include taxes.
The nominal prices domestic petroleum refiners receive for their products have also been
rising much faster than prices received by other U.S. manufacturers. This trend is demonstrated
in Table 3-16 by comparing the producer price index (PPI) for the petroleum refining industry
against the index for all manufacturing industries. Between 1995 and 2006, prices received by
petroleum refineries for their products rose by 223%, while prices received by all manufacturing
firms rose by 26%. The vast majority of this growth in prices has been experienced in the years
after 2002.
Table 3-16. Producer Price Index Industry Data: 1995 to 2006
Petroleum Refining (NAICS 32411)
Total Manufacturing Industries
Year
/
PPI
Annual Percentage
Change in PPI
PPI
Annual Percentage
Change in'PPI
1995
74.5
3%
124.2
3%
1996
85.3
14%
127.1
2%
1997
83.1
-3%
127.5
0%
1998
62.3
-25%
126.2
-1%
1999
73.6
18%
128.3
2%
2000
111.6
52%
133.5
4%
2001
103.1
-8%
134.6
1%
2002
96.3
-7%
133.7
-1%
2003
121.2
26%
137.1
3%
2004
151.5
25%
142.9
4%
2005
205.3
36%
150.8
6%
2006
241.0
17%
156.9
4%
Source: U.S. Bureau of Labor Statistics (BLS). 2007. "Producer Price Index Industry Data: Customizable Industry
Data Tables." Available at . As obtained on October 11,2007.
3.5.5 Profitability of Petroleum Refineries
Estimates of the mean profit (before taxes) to net sales ratios for petroleum refiners are
reported in Table 3-17 for the 2006-2007 fiscal year. These ratios were calculated by Risk
3-32
-------
Table 3-17. Mean Ratios of Profit before Taxes as a Percentage of Net Sales for Petroleum
Refiners, Sorted by Value of Assets
Total 2 Million 10 Million 50 Million 100 Million
Number of Oto 500,000 to to 10 to 50 to 100 to 250 All
Fiscal Year Statements 500,000 2 Million Million Million Million Million Firms
4/1/2006- 44 — — 4.6 6.5 — — 6.7
3/31/2007
Source: Risk Management Association (RMA). 2008. Annual Statement Studies 2007-2008. Pennsylvania: RMA,
Inc.
Management Associates by dividing net income into revenues for 44 firms in the petroleum
refining industry. They are broken down based on the value of assets owned by the reporting
firms.
As these ratios demonstrate, firms that reported a greater value of assets also received a
greater return on sales. For example, firms with assets valued between $10 and $50 million
received a 6.5% average return on net sales, while firms with assets valued between $2 and $10
million only received a 4.6% average return. The average return on sales for the entire industry
was 6.7%.
Obtaining profitability information specifically for small petroleum refining companies
can be difficult as most of these firms are privately owned. However, five of the small, domestic
petroleum refining firms identified in Section 3.4.2.3 are publicly owned companies—the
Arabian American Development Co., CVR Energy Inc., Calumet Specialty Products Partners,
L.P., Holly Corporation, Western Refining, Inc. Profit ratios were calculated for these companies
using data obtained from their publicly available 2006 income statements. These ratios are
presented for in fable 3-18.
3.5.6 Industry Trends
The Energy Information Administration's (ElA's) 2007 Annual Energy Outlook provides
forecasts of average petroleum prices, petroleum product consumption, and petroleum refining
capacity utilization to the year 2030. Trends in these variables are affected by many factors that
are difficult to predict, such as energy prices, U.S. economic growth, advances in technologies,
changes in weather patterns, and future public policy decisions. As a result, the EIA evaluated a
wide variety of cases based on different assumptions of how these factors will behave in the
future. This section focuses on the EIA's "reference case" forecasts, which assume that current
policies affecting the energy sector will remain unchanged throughout the projection period
(EIA, 2007c).
3-33
-------
According to the 2007 Annual Energy Outlook's reference forecast, world oil prices
(defined as the average price of low-sulfur, light crude oil) are expected to fall significantly over
Table 3-18. Net Profit Margins for Publicly Owned, Small Petroleum Refiners: 2006
Net Income
Total Revenue
Net Profit Margin
Company
($millions)
(Smillions)
(%)
Arabian American Development Co.
7.9
98.5
8.0%
Calumet Specialty Products Partners
93.9
1,641.0
5.7%
CVR Energy Inc.
191.6
3,037.6
6.3%
Holly Corporation
266.6
4,023.2
6.6%
Western Refining, Inc.
204.8
4,199.5
4.9%
Sources: Arabian American Development Co. April 6, 2007. 10K for year ended December 31, 2006. EDGAR
Database. Available at .
Calumet Specialty Products Partners. February 23,2007. 10K for year ended December 31,2006. EDGAR
Database. Available at .
CVR Energy Inc. 2006. Google Finance. Available at As
obtained on February 28, 2008.
Holly Corporation. March I, 2007. I OK for year ended December 31, 2006. EDGAR Database. Available at
.
Western Refining, Inc. March 8, 2007. I OK for year ended December 31,2006. EDGAR Database. Available at <
http://www.sec.gov/Archives/edgar/data/1339048/000095013407005096/h44360e 10vk.htm >.
the next 10 years as the amount of oil supplied by non-OPEC and OPEC countries increases.
Since crude oil is the primary input in petroleum refining, a decline in its price would likewise
represent a decline in production costs of petroleum refiners. As a result, the prices of petroleum
products sold to end users are expected to decline over the same period (Table 3-19). These
lower prices will, in turn, encourage more petroleum products to be consumed (Table 3-20).
Between 2007 and 2015, the prices of major petroleum products are expected to fall
approximately 20% to 25%, while consumption of those products is expected to rise by 9%.
Operational capacity of U.S. petroleum refineries is also expected to grow for the foreseeable
future. The expansion of dozens of petroleum refineries has already been announced (Reuters,
2007). The Oil & Gas Journal's 2007 Worldwide Construction Update survey alone catalogued
nearly 40 refining construction projects being pursued in the United States. Table 3-21 lists
selected refinery construction projects that would be subject to the rule. The projects listed
include two refinery expansion projects and five new processes (distillation units and delayed
cokers).
3-34
-------
Table 3-19. Forecasted Average Price of Major Petroleum Products Sold to End Users in
2005 Currency (cents per gallon)
Product
2007
2008
2009
2010
2011
2012
2013
2014
2015
Motor gasoline
257.4
241.3
227.3
217.3
209.2
204.7
201.1
195.2
194.9
Jet fuel
175.4
158.3
152.0
147.2
140.0
135.8
135.5
132.9
133.5
Distillate fuel
253.8
236.6
224.1
215.9
205.0
197.2
194.7
190.3
191.0
Residual fuel oil
123.5
125.8
120.6
113.9
107.7
102.8
96.6
95.9
98.0
LPGs
257.4
241.3
227.3
217.3
209.2
204.7
201.1
195.2
194.9
Source: U.S. Department of Energy, Energy Information Administration (El A). 2007c. "Annual Energy Outlook."
Available at . As obtained on January 21, 2007
Table 3-20. Total Petroleum Products Supplied (millions of barrels per year)
Year
Motor
Gasoline
Jet Fuel
Distillate
Fuel Oil
Residual
Fuel Oil
Liquefied
Petroleum
Gases
Other
Products
Total
2007
3,388
622
1,600
275
819
940
7,643
2008
3,407
646
1,613
278
824
953
7,721
2009
3,446
675
1,631
281
815
955
7,804
2010
3,479
713
1,654
287
809
937
7.879
2011
3,520
728
1,682
289
811
961
7,990
2012
3,563
739
1,710
294
812
958
8,076
2013
3,610
749
1,735
303
812
967
8,177
2014
3,663
758
1,755
306
814
953
8,249
2015
3,716
766
1,774
300
815
970
8,341
Source: U.S. Department of Energy, Energy Information Administration (EIA). 2007c. "Annual Energy Outlook."
Available at . As obtained on January 21, 2007.
Table 3-21. Selected Refinery Construction Projects: 2008-2011
Company and Location
Project
Projected Added
Capacity Expected
(barrels per day) Completion
Cenex Harvest States, Laurel, MT
Frontier Oil Corp, El Dorado, K.S
Marathon Petroleum Co. LLC, Garyville, LA
Motiva Enterprises LLC, Port Arthur, TX
Sinclair Oil Corp, Tulsa, OK
New delayed coker unit N/A
New crude distillation N/A
unit
New vacuum distillation N/A
unit
New crude distillation 180,000
unit
New delayed coker unit N/A
Refinery expansion 325,000
Refinery expansion 45,000
2008
2008
2008
2009
2009
2010
2011
Source: Oil and Gas Journal. November 19, 2007. Worldwide Construction Update.
3-35
-------
In particular, several U.S. refineries are planning projects to expand their ability to handle
cheaper and lower-quality varieties of crude oil (known as "heavy crudes'')- For example,
ConocoPhillips will be expanding its capacity to handle heavy crude oils at its refinery in
Billings, Montana, to 46,000 barrels per day (Reuters, 2007).
In addition to these expansions, two entirely new refineries could potentially be
constructed within the next 5 years. The first is the Arizona Clean Fuels Refinery in Phoenix.
This facility will cost $3 billion to construct and will be capable of producing 6 million gallons
of gasoline, diesel, and jet fuel per day (Arizona Clean Fuels, 2007). Second, a proposal to
construct the MHA Nation Clean Fuels Refinery in North Dakota is being reviewed. If
constructed, this facility will be capable of producing 15,000 barrels of fuel per day (EPA, 2006).
Overall, the EIA forecasts that U.S. operational capacity will increase by a total of 2%
between 2007 and 2015 (Table 3-22). However, since consumption of petroleum products is
projected to grow much more quickly, the rate of capacity utilization is projected to average 90%
during this period.
Table 3-22. Full Production Capacity Utilization Rates for Petroleum Refineries
Year
Petroleum Refineries
Capacity Utilization Rates
(NAICS 324110)
Gross Input to Atmospheric
Crude Oil Distillation Units
(1,000s of barrels per day)
Operational Capacity
(1,000s of barrels per day)
2007
88.8%
15,630
17,597
2008
88.1%
15,587
17,684
2009
88.6%
15,712
17,737
2010
89.1%
15,879
17,822
2011
89.9%
16,055
17,852
2012
90.9%
16,267
17,897
2013
91.4%
16,378
17,914
2014
91.6%
16,433
17,940
2015
92.2%
16,628
18,031
Source: U.S. Department of Energy, Energy Information Administration (EIA). 2007c. "Annual Energy Outlook."
Available at . As obtained on January 21, 2007.
3.6 References
American Petroleum Institute (API). 2006. "FYI on Refineries: Operating at Record Levels."
Available at . As obtained on February 5, 2007.
3-36
-------
Arabian American Development Co. April 6, 2007. 10K for year ended December 31, 2006.
EDGAR Database. Available at .
Calumet Specialty Products Partners. February 23, 2007. 1 OK for year ended December 31.
2006. EDGAR Database. Available at .
CVR Energy Inc. 2006. Google Finance. Available at
As obtained on February 28. 2008.
Dun & Bradstreet Small Business Solutions. Small Business Database. Available at
.
Dun & Bradstreet. 2007. 2007 D&B Million Dollar Directory. Pennsylvania: Dun & Bradstreet
Inc.
Federal Trade Commission (FTC). 2004. "The Petroleum Industry: Mergers, Structural Change,
and Antitrust Enforcement." Available at . As obtained on February 6, 2007.
Gale Research Inc. 2007. Ward's Business Directory of U S Private and Public Companies.
Detroit: Gale Research.
General Accounting Office (FTC). 2007. "Energy Markets: Factors That Influence Gasoline
Prices." Available at . As obtained on
February 6, 2007.
Holly Corporation. March 1, 2007. 10K for year ended December 31, 2006. EDGAR Database.
Available at .
Hoovers. 2007. Free Content, Company Information. Available at
. As obtained on April 11, 2007.
Langinier, Corrine. 2004. "Are Patents Strategic Barriers to Entry?" Journal of Economics and
Business 56:349-361.
MathTech, Inc. 1997. "Industry Profile for the Petroleum Refinery NESHAP." Available at
. As obtained on October
31,2006.
Nicholson, Walter. 2005. Microeconomic Theory: Basic Principles and Extensions. Mason, OH:
South-Western Publishing.
Oil and Gas Journal. 2007. Worldwide Construction Update. November 19, 2007.
3-37
-------
Risk Management Association (RMA). 2008. Annual Statement Studies 2007-2008.
Pennsylvania: RMA, Inc.
Reuters. July 19, 2007. "U.S. Refinery Expansion Plans." Available at
. As obtained on
February 5, 2007.
Small Business Administration. 2006. Table of Small Business Size Standards. Available at
As
obtained on November 30, 2006.
U.S. Bureau of Labor Statistics (BLS). 2007. "Producer Price Index Industry Data: Customizable
Industry Data Tables." Available at . As obtained on October
11,2007.
U.S. Department of Commerce, Bureau of the Census. 1997.1995 Annual Survey of
Manufactures. M95(AS)-I. Washington, DC: Government Printing Office. Available at
. As obtained on October 23,
2006.
U.S. Department of Commerce, Bureau of the Census. 1998.1996 Annual Survey of
Manufactures. M96(AS)-1 (RV). Washington, DC: Government Printing Office.
Available at . As obtained on
October 23, 2006.
U.S. Department of Commerce, Bureau of the Census. 2000. "Population Density by County:
Census 2000 Summary File 1 (SF 1) 100-Percent Data". Available through American
Fact Finder < http://factfinder.census.gov/home/safT/main.html?_lang=en>. As obtained
on February 21, 2008.
U.S. Department of Commerce, Bureau of the Census. 2001.1999 Annual Survey of
Manufactures. M99(AS)-I (RV). Washington, DC: Government Printing Office.
Available at . As obtained on
October 23, 2006.
U.S. Department of Commerce, Bureau of the Census. 2003a. 2001 Annual Survey of
Manufactures. M01(AS)-1. Washington, DC: Government Printing Office. Available at
. As obtained on October 23, 2006.
U.S. Department of Commerce, Bureau of the Census. 2003b. 2001 Annual Survey of
Manufactures. M01(AS)-2. Washington, DC: Government Printing Office. Available at <
http://www.census.gov/prod/2003pubs/m01as-2.pdf. As obtained on March 4, 2008.
U.S. Department of Commerce, Bureau of the Census. 2004. 2002 Economic Census, Industry
Series—Petroleum Refineries. Washington, DC: Government Printing Office. Available
at . As obtained on October 23,
2006.
3-38
-------
U.S. Department of Commerce, Bureau of the Census. 2006. 2005 Annual Survey of
Manufactures. M05(AS)-I. Washington. DC: Government Printing Office. Available at
. As obtained on October 23,
2006.
U.S. Department of Commerce, Bureau of the Census. 2007. 2006 Annual Survey of
Manufactures. Obtained through American Fact Finder Database <
http://factfinder.census.gov/home/saff/main. html?_lang=en>.
U.S. Department of Energy, Energy Information Administration (EIA). "Petroleum Supply
Annuals 1996-2007, Volume 1." Available at
. As obtained on October 31, 2007.
U.S. Department of Energy, Energy Information Administration (EIA). 2006a. "A Primer on
Gasoline Prices." Available at . As obtained on October 31, 2007.
U.S. Department of Energy, Energy Information Administration (EIA). 1999a. "Petroleum: An
Energy Profile 1999." DOE/EIA-0545(99). Available at . As obtained on October 23, 2006.
U.S. Department of Energy, Energy Information Administration (EIA). 2006b. "Refinery
Capacity Report 2006." Available at . As obtained on October 23,
2006.
U.S. Department of Energy, Energy Information Administration (EIA). 2006c. "Petroleum
Supply Annual 2005, Volume 1." Available at . As obtained.on October 31, 2007.
U.S. Department of Energy, Energy Information Administration (EIA). 2007a. "Refinery
Utilization and Capacity." Available at . As obtained on January 7, 2008.
U.S. Department of Energy, Energy Information Administration (EIA). 2007b. "Refiner
Petroleum Product Prices by Sales Type." Available at . As obtained on January 11, 2008.
U.S. Department of Energy, Energy Information Administration (EIA). 2007c. "Annual Energy
Outlook." Available at .
As obtained on January 21, 2007.
U.S. Department of Justice and the Federal Trade Commission. 1992. Horizontal Merger
Guidelines. Available at . As obtained on
October 28,2006.
3-39
-------
U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). 2003.
OSHA Technical Manual, Section IV: Chapter 2, Petroleum Refining Processes. TED 01-
00-015. Washington, DC: U.S. DOL. Available at . As obtained on October 23, 2006.
U.S. Environmental Protection Agency. 1995. "EPA Office of Compliance Sector Notebook
Project: Profile of the Petroleum Refining Industry." EPA/310-R-95-013. Washington,
DC: U.S. EPA. Available at . As obtained on October 23,
2006.
U.S. Environmental Protection Agency. 2006. "Proposed MHA Nation Clean Fuels Refinery."
Available at . As
obtained on February 6, 2007.
Western Refining, Inc. March 8, 2007. 10K for year ended December 31, 2006. EDGAR
Database. Available at .
3-40
-------
/¦
Appendix A. Parent Company Information for Petroleum
Refineries"
Facility Name
City
State
Capacity
(bbl/cd)
Foreign
or
Domestic
Sales
(Smillion)
Employment
Company
Type
(Private or
Public or
Subsidiary)
Owning
Company
Owning
Company
Type
Sales
(Smillion)
Employment (#)
Source
Year of
Data
AGE Refining &
Manufacturing
San Antonio
TX
12,200
D
287
52
Private
D&B
Unknown
Alon USA Energy Inc
Big Spring
TX
67,000
F
Subsidiary
Alon Israel Oil
Company LTD
Private
NA
NA
American Refining
Group
Bradford
PA
10.000
D
350
310
Private
D&B
Unknown
Big West of CA
Bakersfield
CA
66,000
D
Subsidiary
Flying JInc
Private
11.350
16.300
Hoovers
2007
Big West Oil Co
Salt Lake City
UT
29.400
D
S1
Subsidiary
Flying J Inc
Private
11.350
16,300
Hoovers
2007
BP
Whiting
IN
410.000
F
Subsidiary
BP PLC
Public
274,316
97.000
Hoovers
2007
BP
Texas City
TX
437.000
F
Subsidiary
BP PLC
Public
274,316
97,000
Hoovers
2007
BP
Prudhoe Bay
AK
12,500
F
Subsidiary
BP PLC
Public
274.316
97,000
Hoovers
2007
BP
Carson
CA
260.000
F
Subsidiary
BP PLC
Public
274.316
97.000
Hoovers
2007
BP
Ferndale
WA
225,000
F
Subsidiary
BP PLC
Public
274,316
97.000
Hoovers
2007
BP
Toledo
OH
131.000
F
Subsidiary
BP PLC
Public
274,316
97.000
Hoovers
• 2007
Calcasieu Refining Co
Lake Charles
LA
30.000
D
638
51
Private
D&B
Unknown
Calumet Specially
Products
Shreveport
LA
42.000
D
1,641
350
Public
Hoovers
2006
Calumet Specialty
Products
Cotton Valley
LA
13.020
D
1,641
350
Public
Hoovers
2006
Calumet Specialty
Products
Princeton
LA
8.300
D
1.641
350
Public
Hoovers
2006
Cenex Harvest States
Laurel
MT
55.000
D
11.900
6.370
Public
Chevron USA Inc
Perth Amboy
NJ
80.000
D
Subsidiary
Chevron
Corporation
Public
210.1 18
62.500
Hoovers
2006'
Chevron USA Inc
Salt Lake City
UT
45.000.
D
Subsidiary
Chevron
Corpoiation
Public
210.118
62.500
Hoovers
2006
Chevron USA Inc
Portland
OR
D
Subsidiary
Chevion
Corporation
Public
210,118
62,500
Hoovers
2006
Chevron USA Inc
Pascagoula
MS
330,000
D
Subsidiary
Chevron
Corporation
Public
210,118
62,500
Hoovers
2006
(continued)
-------
Appendix A. Parent Company Information for Petroleum
Refineries (continued)
Fiicilily Name
Citv
State
Capacity
(bbl/cd)
Foreign
or
Domestic
Sales
(Smillion)
Employment
Company
Type
(Private or
Public or
Subsidiary)
Owning
Company
Owning
Company
Tvpc
Sales
(Smillion)
¦employment
(«>
Source
Vear of
Data
Chevron USA Inc.
lil Segundo
CA
260.00(1
D
Subsidiary
Chevron
Corporation
Public
210,118
62,500
Hoovers
2006
Chevron USA lnc
Richmond
CA
242,901
D
Subsidiary
Chevron
Corporation
Public
210,118
62,500
Hoovers
2006
Chevron USA lnc
Honolulu (Barbels
Point)
HI
54,000
D
Subsidiary
Chevron
Corporation
Public
210,118
62.500
Hoovers
2006
C'ngo
Corpus Chnsti
TX
156.000
F
Subsidiary
Petrrtleos de
Venezuela S.A
(PDVSA)
Government
Owned
NA
49.180
Hoovers
2004
Citgo Asphalt
Refining Co
Paulsboro
NJ
32.000
F
Subsidiary
PetrAleos de
Venezuela S A
(PDVSA)
Government
Owned
NA
49.180
Hoovers
2004
Citgo Petroleum
Savannah
GA
28,000
¦ F
Subsidiary
PetrAleos de
Venezuela S A
(PDVSA)
Government
Owned
NA
49,180
Hoovers
2004
Citgo Petroleum
Corp
Lake Charles
LA
429.500
F
Subsidiary
PetrAleos de
Venezuela S A
(PDVSA)
Government
Owned
NA
49,180
Hoovers
2004
CofTeyville
Resources LLC
CofTeyville
KS
112,000
D
3,038
577
Public
CVR Energy
lnc
Hoovers
2006
ConocoPhillips
Wesllake
LA
239,400
D
188,523
38.400
Public
Hoovers
2006
ConocoPhillips
Ponca City
OK
194,000
D
188,523
38,400
Public
Hoovers
2006
ConocoPhillips
Billings
MT
58,000
D
188,523
38.400
Public
Hoovers
2006
ConocoPhillips
Borger
TX
146,000
D
188.523
38.400
Public
Hoovers
2006
ConocoPhillips
Sweeny
TX
247,000
D
188,523
38,400
Public
-
Hoovers
2006
ConocoPhillips
Ferndale
WA
' 96,000
D
188,523
38,400
Public
Hoovers
2006
ConocoPhillips
Linden
NJ
238.000
D
188.523
38.400
Public
Hoovers
2006
ConocoPhillips
Wood River
LA-
Carson/Wilmington
IL
306.000
D
188,523
38,400
Public
Hoovers
2006
ConocoPhillips
CA
139,000
D
188,523
38,400
Public
Hoovers
2006
(continued)
-------
Appendix A. Parent Company Information for Petroleum
Refineries (continued)
Facility Name
Citv
State
Capacity
(bbl/cdj
Foreign
or
Domestic
Sales
(Smillion)
Employment
Company
Type
(Private or
Public or
Subsidiary)
Owning
Company
Owning
Company
Type
Sales
(Smillion)
Employment
m
Source
Year of
Data
ConocoPhillips
SF - Rodeo
CA
76.000
D
188,523
38,400
Public
Hoovers
2006
ConocoPhillips
Arroyo Grande
(Santa Maria)
CA
44,200
D
188,523
38.400
Public
Hoovers
2006
ConocoPhillips
Belle Cliasse
LA
247,000
D
188,523
38,400
Public
Hoovers
2006
ConocoPhillips
Trainer (Marcus
Hook)
PA
185,000
D
188,523
38.400
Public
Hoovers
2006
ConocoPhillips
Kuparuk
AK
14,000
D
188.523
38,400
Public
Hoovers
2006
Countrymark
Cooperative, Inc
Mt Vernon
IN
23,000
D
87
300
Private
Cross Oil & Refining
Co Inc
Smackover
AR
7.200
D
49
no
Private
Delek Refining Ltd
Tyler
TX
58.000
F
Subsidiary
Delek Group
LTD
Public
6.237
2,803
Hoovers
2006
Edgington Oil Co
Long Beach
CA
26.000
F
Subsidiary
Alon Israel Oil
Company LTD
Private
NA
NA
Ergon Refining Inc
Vicksburg
MS
23.000
D
Subsidiary
Ergon.Inc
Private
1.300
2.300
Ergon-Wesl Virginia
Inc
Newell (Congo)
WV
20,000
D
Subsidiary
Ergon, Inc
Private
1,300
2,300
ExxonMobil Corp
Baton Rouge
LA
501,000
D
377,635
82,100
Public
Hoovers
2006
ExxonMobil Corp
Billings
MT
60.000
D
377,635
82,100
Public
Hoovers
2006
ExxonMobil Corp
Joliet
IL
238.500
D
377,635
82,100
Public
Hoovers
2006
ExxonMobil Corp
Beaumont
TX
348,500
D
377,635
82,100
Public
Hoovers
2006
ExxonMobil Corp
Torrance
CA
149,500
D
377,635
82,100
Public
Hoovers
2006
ExxonMobil Corp
Chalmette
LA
188,160
D
377,635
82,100
Public
Hoovers
2006
ExxonMobil Oil Corp
Baylown
TX
562.500
D
Subsidiary
ExxonMobil
Corp
Public
377,635
82.100
Hoovers
2006
Flint Hills Resources
Corpus Chrisli
TX
288.126
D
Subsidiary
Koch Industries
Inc
Private
"51.500
85.000
Flint Hills Resources
North Pole
AK
210.0(10
D
Subsidiary
Koch Industries
Inc
Private
51.500
85,000
(continued)
-------
Appendix A. Parent Company Information for Petroleum Refineries (continued)
Company
Type
Foreign
(Private or
Owning
Capacity
or
Sales
Public or
Owning
Company
Sales
Employment
Year of
Fncilitv Name
C.'itv
Slate
(bbl/cd)
Domestic
(Smillinn)
Employment
Subsidiary)
Company
Tvpe
(Smillion)
w
Source
Data
Koch Industries
Flint Hills Resources
Rosemount
Tonopah/Eagle
MN
279,300
D
Subsidiary
Inc
Private
51,500
85.000
Foreland Refining Co
Springs
NV
2.000
P
56
100
Private
D&B
Unknown
Frontier Oil & Refining
Co
Cheyenne
WY
47.000
D
Subsidiary
Frontier Oil Corp
Private
4.000
727
Frontier Oil Corp
El Dorado
KS
106,000
D
4.000
727
Private
Western Refining,
Giant Refining Co
Yorktown
VA
58.600
D
Subsidiary
Inc
Western Refining,
Private
4.200
416
Hoovers
2006
Giant Refining Co
Bloomfield
NM
16,800
D
Subsidiary
Inc
Western Refining,
Private
4.200
416
Hoovers
2006
Giant Refining Co
Gallup
NM
20,800
D
Subsidiary
Inc
Private
4.200
416
Hoovers
2006
Goodwaj Refining LLC
Atmore
AL
4,100
D
3
18
Private
D&B
Unknown
Greka Integrated Inc
Santa Maria
CA
9,500
D
22
145
Private
Gulf Atlantic
-
-
Operations LLC
Mobile Bay
AL
16.700
D
9
32
Private
D&B
Unknown
Hess Corporation
Port Reading
NJ
D
23,200
11.610
Public
Holly Corp
Woods Cross
UT
24.700
D
4.023
859
Public
Hoovers
2006
Hunt Refining Co
Tuscaloosa
AL
34.500
D
4,871
1.100
Private
Ward's
2007
Hunt Southland
Hunt Refining
Refining
Lumberton
MS
D
Subsidiary
Co
Private
4.871
1,100
Ward's
2007
Hunt Southland
Hunt Refining
Refining
Sandersville
MS
11,000
D
Subsidiary
Co
Private
4.871
1.100
Ward's
2007
Kern Oil & Refining
Co
Bakersfield
CA
26.000
D
NA
NA
Private
Lion Oil Co
El Dorado
AR
70.000
D
247^
425
Private
Little America Refining
Evansville
-
Sinclair
Co
(Casper)
WY
24.500
D
Subsidiary
Companies
Private
5.500
7.000
Lunday-Thagard Co
South Gate
CA
8,500
D
Subsidiary
World Oil Corp
Private
277
475
Hoovers
2007
(continued)
-------
Appendix A. Parent Company Information for Petroleum Refineries (continued)
Facility Name
City
State
Capacity
(bbl/cdj
Foreign
or
Domestic
Sales
(Smillion)
Employment
Company
Type
(Private or
Public or
Subsidiary)
Owninj;
Company
Owning
Company
Type
Sales
(Smillion)
Employment
(#>
Source
Year
of
Data
Lyondell-Citgo
Refining Co
Houston
TX
270.200
D
Subsidiary
Lyondcll
Chemical Co
Public
18.600
10,880
Marathon Petroleum
Co LLC
Robinson
II.
192.000
D
Subsidiary
Marathon Oil
Corp
Public
65,449
28,195
Hoovers
2006
Maralhon Petroleum
Co LLC
Catlettshurg
KY
222.000
D
Subsidiary
Marathon Oil
Corp
Public
65.449
28.195
Hoovers
2006
Marathon Petroleum
Co LLC
Detroit
Ml
100,000
f)
Subsidiary
Marathon Oil
Corp
Public
65.449
28,195
Hoovers
2006
Marathon Petroleum
Co LLC
Canton
OH
73.000
D
Subsidiary
Marathon Oil
Corp
Public
65.449
28.195
Hoovers
2006
Marathon Petroleum
Co LLC
St Paul Park
MN
70.000
D
Subsidiary
Marathon Oil
Corp
Public
65.449
28.195
Hoovers
2006
Marathon Petroleum
Co LLC
Texas City
TX
72.000
D
Subsidiary
Marathon Oil
Corp
Public
65,449
28,195
Hoovers
2006
Marathon Petroleum
Co LLC
Garyville
LA
245.000
D
Subsidiary
Marathon Oil
Corp
Public
65,449
28,195
Hoovers
2006
Montana Refining Co
Great Falls
MT
8,200
F
Subsidiary
Connacher Oil
and Gas Limited
Public
NA
NA
Moliva Enterprises
Norco
LA
226.500
D
32,100
2,700
Private
Motiva Enterprises
Port Arthur
TX
285,000
D
32,100
2,700
Private
Motiva Enterprises
Convent
LA
235,000
D
32,100
2,700
Private
Murphy Oil USA Inc
Superior
WI
34,300
D
Subsidiary
Murphy Oil Corp
Public
14,307
7,296
Hoovers
2006
Murphy Oil USA Inc
Meraux
LA
120,000
D
Subsidiary
Murphy Oil Corp
Public
14,307
7.296
Hoovers
2006
National Cooperative
Refinery Association
McPherson
K.S
81,200
D
Subsidiary
CcneN Harvest
States
Public
1 1,900
6,370
Navajo Refining Co
Artesia
NM
75,000
D
Subsidiary
Holly Corp
Public
4,023
859
Hoovers
2006
Paramount Petroleum
Corp
Paramount
CA
50,000
F
Subsidiary
Alon Israel Oil
Company LTD
Private
NA
NA
Pasadena Refining
Systems Inc
Pasadena
TX
100,000
F
Subsidiary
Petroleo
Brasileiro. S A
Government
Owned
72,347
62.266
Hoovers
2006
(continued)
-------
Appendix A. Parent Company Information for Petroleum Refineries (continued)
Facility Name
C.'ilv
State
Capacity
(bbl/cd)
Foreign
or
Domestic
Sales
(Smillion)
Employment
Compan)
Type
(Private or
Public or
Subsidiary)
Owning
Company
Owning
Company
Tvpe
Sales
(Smillion)
Employment
(»)
Sou rce
Year of
Data
PDV Midwest Refining
Lernont
IL
167.0(10
F
Subsidiary
Petroleos de
Venezuela S A
(PDVSA)
Government
Owned
NA
NA
Pelican Refining Co
LLC
Lake Charles
LA
D
29
62
Private
Petro Star Inc
North Pole
AK
17,000
D
Subsidiary
Arctic Slope
Regional Corp
Private
1,500
5.743
Petro Star Inc
Valdez
AK
48,000
D
Subsidiary
Arctic Slope
Regional Corp
Private
1,500
5,743
Placid Refining Inc.
Port Allen
LA
56.000
D
1,400
200
Private
San Joaquin Refining
Co . Inc
Bakers field
CA
15.000
D
288
20
Private
Shell Chemical LP
St Rose
LA
55.000
F
Subsidiary
Royal Dutch
Shell. PLC
Public
312,323
108,000
Hoovers
2006
Shell Chemical LP
Snraland
AL
80.000
F
Subsidiary
Royal Dutch
Shell. PLC
Public
312,323
108,000
Hoovers
2006
Shell Oil Products US
Anacorlcs
WA
145.000
F
Subsidiary
Royal Diuch
Shell. PLC
Public
312,323
108,000
Hoovers
2006
Shell Oil Products US
Martinez
CA
155.600
F
Subsidiary
Royal Dulch
Shell. PLC
Public
312.323
108.000
Hoovers
2006
Shell Oil Products US
Wilmington
CA
98.500
F
Subsidiary
Royal Dulch
Shell. PLC
Public
' 312,323
108.000
Hoovers
2006
Shell Oil Products US -
Deer Park Refining
Limited Partnership
Deer Park
TX
333.700
F
Subsidiary
Royal Dulch
Shell. PLC
Public
312,323
108,000
Hoovers
2006
Silver Eagle Refining
Inc
Evanston
WY
3.000
D
NA
NA
Private
Silver Eagle Refining
Inc
Woods Cross
UT
10,250
D
NA
NA
Private
Sinclair Oil Corp
Tulsa ¦
OK
70.300
D
'Subsidiary
Sinclair
Companies
Private
5,500
7,000
(continued)
-------
Appendix A. Parent Company Information for Petroleum Refineries (continued)
Fucilitv Name
Citv
State
Capacity
(hbl/cd)
Foreign
or
Domestic
Sales
(Smillion)
Employment
Company
Type
(Private or
Public or
Subsidiary)
Owning
Company
Owning
Company
Tvpc
Sales
(Smillion)
Employment (W)
Source
Year of
Data
Sinclair Oil Corp
Sinclair
WY
66.000
D
Subsidiary
Sinclair
Companies
Private
5.500
7.000
Somerset Refinery Inc
Somerset
KY
5.500
D
Subsidiary
Somerset Oil Inc
Private
55
150
South Hampton
ResourcesInc
Silsbee
TX
D
Subsidiary
Arabian
American
Development Co
Public
80
118
Suncor Energy
Commerce City
CO
62.000
F
Subsidiary
Suncor Energy
Inc
Public
13,583
5,152
Hoovers
2006
Suncor Energy
Denver
CO
32.000
F
Subsidiary
Suncor Energy
Inc
Public
13.583
5.152
Hoovers
2006
Sunoco. Inc
Westville
NJ
145.000
D
38.715
14,000
Public
Hoovers
, 2006
Sunoco, Inc
Marcus Hook
PA
1 75.000
D
38,715
14.000
Public
Hoovers
2006
Sunoco.Inc
Toledo
OH
160,000
D
38,715
14,000
Public
Hoovers
2006
Sunoco.Inc
Tulsa
OK
85.000
D
38.715
14.000
Public
Hoovers
2006
Sunoco. Inc
Phil. (Girard Pt &
Pt Breeze)
PA
335,000
D
38,715
14,000
Public
Hoovers
2006
Ten ByInc
Oxnard
CA
2,800
NA
NA
Tesoro
Mandan
ND
58,000
D
Subsidiary
Tesoro Corp
Public
18.104
3,950
Hoovers
2006
Tesoro
Salt Lake City
UT
58,000
D
Subsidiary
Tesoro Corp
Public
18,104
3,950
Hoovers
2006
Tesoro
Anac'ortes
WA
120,000
D
Subsidiary
Tesoro Corp
Public
18,104
3,950
Hoovers
2006
Tesoro
Golden Eagle
CA
166.000
D
Subsidiary
Tesoro Corp
Public
18.104
3,950
Hoovers
2006
T esoro
Kapolei
HI
93,500
D
Subsidiary
Tesoro Corp
Public
18.104
3,950
Hoovers
2006
T esoro
Kenai
AK
72.000
D
Subsidiary
Tesoro Corp
Public
18.104
3,950
Hoovers
2006
Total SA
Port Arthur
TX
232,000
F
175,189
95,070
Public
Hoovers
2005
Trigeant Ltd
Corpus Christi
TX
D
5
50
Private
D&B
Unknown
United Refining Co
Wairen
PA
65,000
D
Subsidiary
Red Apple Group
Inc
Private
4.200
7.000
US Oil & Refining Co
Tacoma
WA
37,850
NA
NA
Valero Energy
Corpus Christi
TX
142,000
D
91,833
21,836
Public
Hoovers
2006
(continued)
-------
Appendix A. Parent Company Information for Petroleum Refineries (continued)
Facilitv Name
Citv
Stale
Capacity
(bbl/cil)
Foreign
or
Domestic
Sales
(Smillion)
Employment
Company
Type
(Private or
Public or
Subsidiary)
Owning
Company
Owning
Company
Type
Sales
(Smillion)
Employment (#)
Source
Year of
Data
Valero Energy
Houston
TX
83.00(1
D
91.833
21,836
Public
Hoovers
2006
Valero Energy
Texas City
TX
213,750
D
91.833
21.836
Public
Hoovers
2006
Valero Energy
Krotz Springs
LA
80.000
D
91,833
21.836
Public
Hoovers
2006
Valero Energy
Bemcia
CA
144,000
D
91,833
21.836
Public
Hoovers
2006
Valero Energy
Wilmington
CA
6,200
D
91,833
21.836
Public
Hoovers
2006
Valero Energy
Norco
LA
185.003
D
91.833
21.836
Public
Hoovers
2006
Valero Energy
Delaware City
DE
181.500
D
91.833
21,836
Public
Hoovers
2006
Valero Energy
Lima
OH
146,900
D
91,833
21,836
Public
Hoovers
2006
Valero Energy
Memphis
TN ,
1 80,000
D
91.833
21.836
Public
Hoovers
2006
Valero Energy
Three Rivers
TX
90.000
D
91.833
21,836
Public
Hoovers
2006
Valero Energy
Sunray
TX
158,327
D
91,833
21.836
Public
Hoovers
2006
Valero Energy
Ardmore
OK
83.640
D
91.833
21.836
Public
Hoovers
2006
Valero Energy
Wilmington
CA
80.887
D
91.833
21,836
Public
Hoovers
2006
Valero Energy
Paulsboro
NJ
160,000
D
91,833
21,836
Public
Hoovers
2006
Valero Energy
Port Arthur
TX
260,000
D
91,833
21,836
Public
Hoovers
2006
Western Refining, Inc
El Paso
TX
1 16,000
D
4,200
416
Public
Hoovers
2006
Wynnewood Refining
Co
W> nnewood
OK
54,000
D
97
200
Subsidiary
Gary-Williams
Co
Private
Note: All data were collected from the 2007 D&B Million Dollar Direction unless noted other wise. Data collected from the 2006 D&B Small Business Database
are indicated using "D&B" in the source column. Data collected from Ward's Business Directory are identified using "Ward's" in the source column.
"These data are shown with the permission of D&B.
Sources: Dun & Bradstreet. 2007. 2007 D&B Million Dollar Directory. Pennsylvania: Dun & Bradstreet Inc.
Dun & Bradstreet Small Business Solutions. Small Business Database. Available at .
Gale Research Inc. 2007. Ward's Business Directory of U S Private and Public Companies. Detroit: Gale Research.
Hoovers. 2007. Free Content, Company Information. Available at . As obtained on April 11, 2007.
-------
SECTION 4
NSPS REGULATORY OPTIONS, COSTS AND EMISSION REDUCTIONS FROM
COMPLYING WITH THE NSPS
This section of the RIA provides descriptions of the regulatory options considered in the
developing of this final NSPS. and also the costs and emission reductions estimated for each
option. An appendix to this section (Appendix B) provides details on the rationale behind the
choice of each option that is included in this NSPS.
4.1 Background Information on the Setting of NSPS
New source performance standards (NSPS) implement 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. As a result of our periodic review of the NSPS for petroleum refineries (40 CFR part
60, subpart J), we proposed amendments to the current standards of performance and separate
standards of performance for new process units (72 FR 27278, May 14, 2007). In response to
several requests, we extended the 60-day comment period from July 13, 2007, to August 27,
2007 (72 FR 35375, June 28, 2007). We also issued a notice of data availability (72 FR 69175,
December 7, 2007) (NODA) to notify the public that additional information had been added to
the docket; the NODA also extended the public comment period on the proposed rule to January
7, 2008. We received a total of 38 comments from refineries, industry trade associations, and
consultants; State and local environmental and public health agencies; environmental groups;
and members of the public during the extended comment period and 8 additional comments on
4-1
-------
the NODA. These final rules reflect our full consideration of all of the comments we received.
Detailed responses to the comments are contained in the Response to Comments document
which is included in the docket for this rulemaking.
4.2 Summary of the Final NSPS and Changes Since Proposal
We are promulgating 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 final standards in 40 CFR part 60, subpart Ja. For example, we are revising
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.
Consequently, these vapors are exempt from the sulfur dioxide (SO2) treatment standard in 40
CFR 60.104(a)(1) and are not required to be monitored. In a related amendment, we are
clarifying that monitoring is not required for fuel gases that are identified as inherently low
sulfur or demonstrated to contain a low sulfur content. We are also revising the coke bum-off
equation to account for oxygen (02)-enriched air streams. Other amendments include
clarification of definitions and correction of grammatical and typographical errors.
.J The final standards in 40 CFR part 60, subpart Ja include emission limits for fluid
catalytic cracking units (FCCU), fluid coking units (FCU), sulfur recovery plants (SRP), and fuel
gas combustion devices. Subpart Ja also includes work practice standards for minimizing
emissions of volatile organic compounds (VOC) from flares and S02 emissions from fuel gas
combustion devices and for reducing emissions of VOC from delayed coking units. Only those
affected facilities that begin construction, modification, or reconstruction after May .14, 2007 will
be affected by the standards in 40 CFR part 60, subpart Ja. Units for which construction,
modification, or reconstruction began on or before May 14, 2007 must continue to comply with
the applicable standards under the current NSPS in 40 CFR part 60, subpart J, as amended.
4.2.1 Final Amendments to the NSPS
As proposed, we are amending the definition oP'fuel gas" to specifically exclude 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 are still considered to be affected fuel gas combustion devices if they
combust other gases that meet the definition of fuel gas. and all auxiliary fuel gas fired to these
devices are subject to the fuel gas limit; however, continuous monitoring is not required for the
vapors collected from wastewater or marine vessel loading operations that are being incinerated
4-2
-------
because these gases are not considered to be fuel gases under the definition of "fuel gas" in
subpart J.
We are also exempting certain fuel gas streams from all continuous monitoring
requirements. Monitoring is currently not required for combustion in a flare of process upset
gases or flaring of gases from relief valve leakage or emergency malfunctions. Additionally,
monitoring is not 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
of 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 are required to document the -
exemption for which each fuel gas stream applies and ensure that the stream remains qualified
for that exemption.
We proposed to amend the definition of "Claus sulfur recovery plant" in 40 CFR
60.101 (i) to clarify that the SRP may consist of multiple units and that primary sulfur pits are
considered part of the Claus SRP based on a recent applicability determination. However, due to
concerns regarding retroactive non-compliance, we are not amending this definition in the final
amendments for subpart J. Similarly, we proposed revisions to the subpart J definitions of
"oxidation control system" and "reduction control system" in 40 CFR 60.101 (j) and 40 CFR
60.101 (k), respectively, to clarify that these systems were intended to recycle the sulfur back to
the Claus SRP. The proposed amendments needlessly limit the types of tail gas treatment
systems that can be used; therefore, we are not amending these definitions in the final
amendments for subpart J.
4.2.2 Final requirements for New FCCU and New FCU
The final standards for new fluid catalytic cracking units include emission limits for
particulate matter (PM), SO2, nitrogen oxides (NOx), and carbon monoxide (CO). The final
standards include no universal opacity limit because the opacity limit in subpart J is intended to
ensure compliance with the PM limit. Subpart Ja requires that sources use direct PM monitoring,
bag leak detection systems, or parameter monitoring (along with annual emission tests) to ensure
compliance with the PM limit. A provision for a site-specific opacity operating limit is provided
for units that meet the PM emission limits using a cyclone.
For PM emissions from new FCCU and new FCU, we proposed a PM limit of 0.5 pounds
(lb)/l ,000 lb coke burnoffin the regenerator or (if a PM continuous emission monitoring system
(CEMS) is used), 0.020 grains per dry standard cubic feet (gr/dscf) corrected to 0% excess air.
4-3
-------
We have revised the final PM standards to establish separate limits for modified or reconstructed
FCCU (1 lb/1,000 lb coke burn or 0.040 gr/dscf corrected to 0% excess air) and newly
constructed FCCU (0.5 lb/1,000 lb coke burn or 0.020 gr/dscf corrected to 0% excess air). The
final PM limit for new, modified, or reconstructed FCU is 1 lb/1,000 lb coke burn or 0.040
gr/dscf corrected to 0% excess air).
Initial compliance with the PM emission limits for FCCU and FCU is determined using
EPA Method 5, 5B or 5F (40 CFR part 60, appendix A) instead of being restricted to only EPA
Method 5 as previously proposed. Procedures for computing the PM emission rate using the total
PM concentration, effluent gas flow rate, and coke burn-off rate are the same as in 40 CFR part
60, subpart J, as amended. To demonstrate ongoing compliance, an owner or operator must
monitor PM emission control device operating parameters and conduct annual PM performance
tests, use a PM CEMS. or operate bag leak detection systems and conduct annual PM
performance tests. A new alternative allows refineries with wet scrubbers as PM control devices
to use the approved alternative in 40 CFR 63.1573(a) for determining exhaust gas flow rate
instead of a continuous parameter monitoring system (CPMS). An alternative to the requirements
for monitoring the pressure drop from wet scrubbers that are equipped with jet ejectors or
atomizing spray nozzles is to conduct a daily check of the air or water pressure to the nozzles and
record the results of each inspection. The final rule also includes procedures for establishing an
alternative opacity operating limit for refinery that use continuous opacity monitoring systems
,/
(COMS); this alternative is allowed only for units that choose to comply with the PM limit using
cyclones. If operating parameters are 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 three-run average of the values for the applicable parameters measured over the three test
runs. If ongoing compliance is demonstrated using a PM CEMS, the CEMS must meet the
conditions in Performance Specification 11 (40 CFR part 60, appendix B) and the quality
assurance (QA) procedures in Procedure 2, 40 CFR part 60, appendix F. The relative response
audits must be conducted annually (in lieu of annual performance tests for units not employing a
PM CEMS) and response correlation audits must be conducted once every 5 years.
For NOx emissions from affected FCCU and FCU, we proposed a limit of 80 ppmv based
on a 7-day rolling average (dry basis corrected to 0% excess air) and co-proposed having no limit
for FCU. We are adopting the 80 ppmv NOx emission limits for FCCU and FCU as proposed.
Initial compliance with the 80 ppmv emission limit is demonstrated by conducting a performance
evaluation of the CEMS in accordance with Performance Specification 2 in 40 CFR part 60.
4-4
-------
appendix B, with Method 7 (40 CFR part 60. appendix A) as the reference method. Ongoing
compliance with these emission limits is determined using the CEMS to measure NOx emissions
as discharged to the atmosphere, averaged over 7-day periods.
No changes have been made to the proposed SO2 emission-limits for affected FCCU and
FCU. The final SO2 emission limits are to maintain SO2 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% moisture and 0% excess air). Initial
compliance with the final SO2 emission limits is demonstrated by conducting a performance
evaluation of the SO2 CEMS in accordance with Performance Specification 2 in 40 CFR part 60,
appendix B with Method 6, 6A, or 6C of 40 CFR part 60, appendix A as the reference method.
Ongoing compliance with both SO2 emission limits is determined using the CEMS to measure
SO2 emissions as discharged to the atmosphere, averaged over the 7-day and 365-day averaging
periods.
No changes have been made since proposal to the CO limits. The final CO emission limit
for affected fluid catalytic cracking units and FCU is 500 ppmv ( I-hour average, dry at 0%
excess air). Initial compliance with this emission limit is demonstrated by conducting a
performance evaluation for the CEMS in accordance with Performance Specification 4 in 40
CFR part 60, appendix B, with Method 10 or 1 OA 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 is determined on an hourly basis using the CEMS to measure
CO emissions as discharged to the atmosphere. An exemption from monitoring may be requested
for an FCCU or FCU 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 FCCU. As proposed, units that are exempted from the CO monitoring
requirements must comply with control device operating parameter limits.
4.2.3 Final Requirements for New Sulfur Recovery Plants (SRP)
For new, modified, and reconstructed SRP with a capacity greater than 20 long tons per
day (LTD), we proposed a limit of 250 ppmv total sulfur (combined SO2 and reduced sulfur
compounds) as SO2 (dry basis at 0% excess air determined on a 12-hour rolling average basis).
The refinery could comply with the limit for each process train or release point or with a flow
rate weighted average of 250 ppmv for all release points. For affected SRP with a capacity less
than 20 LTD, we proposed a mass emissions limit for total sulfur equal to 1 weight percent or
less of sulfur recovered (determined hourly on a 12-hour rolling average basis).
4-5
-------
In this final rule, we are adopting the current limits in subpart J (which include separate
emission limits for oxidative and reductive systems) for affected SRP with a capacity greater
than 20 LTD. For these affected SRP, the final limits for SRP having an oxidation control system
or a reduction control system followed by incineration is 250 ppmv (dry basis) of SO2 at 0%
excess air. For an affected SRP with a reduction control system not followed by incineration, the
final limit is 300 ppmv of reduced sulfur compounds and 10 ppmv of hydrogen sulfide (H2S),
each calculated as ppm SO2 by volume (dry basis) at 0% excess air. If the SRP consists of
multiple process trains or release points, the limits apply to each process train or release points.
A new alternative allows refineries to use a correlation to calculate their effective emission limit
for Claus SRP that use oxygen enrichment in the Claus burner. For an affected SRP with a
capacity of 20 LTD or less, the sulfur recovery efficiency standard is based on a sulfur recovery
efficiency of 99%. However, due to the difficulties associated with on-going monitoring of SRP
recovery efficiency, in this final rule, we are promulgating concentration limits that correlate
with a sulfur recovery efficiency of 99%. For a Claus unit with an oxidative control system or
any small SRP followed by an incinerator the emission limit is 2,500 ppmv (dry basis) of SO2 at
0% excess air. For all other small SRP, the emission limit is 3,000 ppmv reduced sulfur
compound and 100 ppmv H2S, each calculated as ppm SO2 by volume (dry basis) at 0% excess
air. A similar correlation is provided for small Claus SRP that use oxygen enrichment, similar to
that provided for large SRP. The standards for small SRP apply to all release points from the
SRP combined (note that secondary sulfur storage units are not considered part of the SRP). We
are not promulgating the H2S limit of 10 ppmv (dry basis, at 0% excess air determined on a 12-
hour rolling average basis) or related operating limits that were included in §60.102a(e) and (f)
of the proposed rule.
Initial compliance with the emission limit for large SRP (capacity greater than 20 LTD) is
demonstrated by conducting a performance evaluation for the SO2 CEMS in accordance with
either Performance Specification 2 (40 CFR part 60, appendix B) for SRP with oxidation control
systems or reduction control systems followed by incineration or Performance Specification 5
(40 CFR part 60, appendix B) for SRP with reduction control systems not followed by
incineration.
Ongoing compliance with the SO2 limits for large SRP is determined using an SO2
CEMS (for oxidative or reductive systems followed by incineration) or a CEMS that uses an air
or O2 dilution and oxidation system to convert the reduced sulfur to SO2 and then measures the
total resultant SO2 concentration (for reductive systems not followed by incineration). An O2
4-6
-------
monitor is also required for converting the measured combined SO2 concentration to the
concentration at 0% O2.
Initial and ongoing compliance requirements for small SRPs are the same as for large
SRPs.
4.2.4 Final Requirements for New Fuel Gas Combustion Devices .
In the subpart Ja proposal, we divided fuel gas combustion units into two separate
affected sources: "process heaters" and "other fuel gas combustion devices." In response to
comments, we have eliminated the proposed definition of "other fuel gas combustion devices"
and revised the standards to either refer to fuel gas combustion devices, which include process
heaters, or to refer specifically to process heaters. This revision makes the definition of "fuel gas
combustion devices" consistent with subpart J. We have also added a definition of "flare."
We proposed a primary sulfur oxides emission limit for fuel gas combustion devices of
20 ppmv or less SO2 (dry at 0% excess air) on a 3-hour rolling average basis and 8 ppmv or less
on a 365-day rolling average basis. We also proposed an alternative limit of 160 ppmv H2S or, in
the case of coker-derived fuel gas, for total reduced sulfur (TRS).on a 3-hour rolling average
basis and 60 ppmv or less on a 365-day rolling average basis. We are promulgating the 20 ppmv
and 8 ppmv limits for SO2 as proposed. We are also promulgating the alternative limit, except
that the limits are expressed and measured as H2S in all cases. The alternative H2S limit is 162
ppmv or less in the fuel gas on a 3-hour rolling average basis and 60 ppmv or less in the fuel gas
on a 365-day rolling average basis. The final rule does not include an alternative TRS limit for
S02.
Initial compliance with the 20 ppmv SO2 limit or the 162 ppmv H2S concentration limits
is demonstrated by conducting a performance evaluation for the CEMS. The performance
evaluation for an SO2 CEMS is conducted in accordance with Performance Specification 2 in 40
CFR part 60, appendix B. The performance evaluation for an H2S CEMS is conducted in
accordance with Performance Specification 7 in 40 CFR part 60, appendix B. Ongoing
compliance with the sulfur oxides emission limits is determined using the applicable CEMS to
measure either SO2 in the exhaust gas to the atmosphere or H2S in the fuel gas, averaged over the
3-hour and 365-day averaging periods.
Similar to clarifications for 40 CFR part 60, subpart J, the definition of "fuel gas"
includes exemptions for vapors collected and combusted in an air pollution control device
installed to comply with specified wastewater or marine vessel loading provisions. We are also
4-7
-------
streamlining the process for an owner or operator to demonstrate that a fuel gas stream not
explicitly exempted from continuous monitoring is inherently low sulfur.
/ '
For new, modified, or reconstructed process heaters with a rated capacity greater than 20
million British thermal units per hour (MMBtu/hr), we proposed a NOx limit of 80 ppmv (dry
basis, corrected to 0% excess air) on a 24-hour rolling average basis. The final NOx emission
limit for affected process heaters is 40 ppmv on a 24-hour rolling average basis (dry at 0%
excess air) for process heaters greater than 40 MMBtu/hr. For process heaters greater than 100
MMBtu/hr capacity, initial compliance with the 40 ppmv emission limit is demonstrated by
conducting a performance evaluation of the CEMS in accordance with Performance
Specification 2 in 40 CFR part 60, appendix B. For process heaters between 40 MMBtu/hr and
100 MMBtu/hr capacity, initial compliance is demonstrated using EPA Method 7. For process
heaters greater than 100 MMBtu/hr capacity, ongoing compliance with this emission limit is
determined using the CEMS to measure NOx emissions as discharged to the atmosphere,
averaged over 24-hour periods. For process heaters between 40 MMBtu/hr and 100 MMBtu/hr
capacity, ongoing compliance with this emission limit is determined using biennial performance
tests.
4.2.5 Final Work Practice Standards
We proposed three work practice standards to reduce SCb, VOC, and NOx emissions
from flares and from startup, shutdown, and malfunction events and to reduce VOC and SO2
emissions from delayed coking units. We also co-proposed to require only one of these work
practice standards: the requirement to depressure delayed coking units. This proposed standard
required new delayed coking units to depressure to 5 pounds per square inch gauge (psig) during
reactor vessel depressuring and vent the exhaust gases to the fuel gas system.
We are promulgating a work practice standard for delayed coking units and modified
requirements to reduce emissions from flares. The final work practice standard for delayed
cokers requires affected delayed coking units to depressure to 5 pounds per square inch gauge
(psig) during reactor vessel depressuring. We are requiring the exhaust gases to be vented to the
fuel gas system as proposed or to a flare.
To reduce SO2 emissions from the combustion of sour fuel gases, the final rule requires
refineries to conduct a root cause analysis of any emissions limit exceedance or process start-up,
shutdown, upset, or malfunction that causes a discharge into the atmosphere, either directly or
indirectly, from any fuel gas combustion unit subject to the provisions of subpart Ja that exceeds
4-8
-------
500 pounds per day (lb/day) of SO2. Recordkeeping and reporting requirements apply in the
event of such a discharge.
We are not promulgating the proposed definition of "fuel gas producing unit" and the
proposed requirement for "no routine flaring." Instead, we are promulgating the following
requirements: (1) flare fuel gas flow rate monitoring; (2) a flare fuel gas flow rate limit; and (3) a
flare management plan. Affected flares cannot exceed 250,000 standard cubic feet per day (scfd)
on a 30-day rolling average basis. In cases where the flow exceeds this value, this would require
installation of a flare gas recovery system or other methods to reduce flaring from the affected
flare. A provision is provided for an exclusion from the flow limitation for times when the
refinery can demonstrate that the refinery produces more fuel gas than it needs to fuel the
refinery combustion devices (i.e., it is fuel gas rich) or that the flow is due to an upset or.
malfunction, provided the refinery follows procedures outlined in the flare management plan.
The flare management plan should address potential causes of fuel gas imbalances (i.e.. excess
fuel gas) and records to be maintained to document these periods. To demonstrate compliance
with the flow limitations, flow rate monitors must be installed and operated. As described in 40
CFR 60.103a(a), the flare management plan must include a diagram illustrating all connections
to each affected flare, identification of the flow rate monitoring device and a detailed description
of the manufacturer's specifications regarding quality assurance procedures, procedures to
. maintain to minimize flaring during planned start-up and shut down events, and procedures for
implementing root cause analysis when daily flow to the flare exceeds 500,000 scfd. The root
cause analysis procedures should address the evaluation of potential causes of upsets or
malfunctions and records to be maintained to document the cause of the upset or malfunction.
Excess emission events for the flow rate limit of 250,000 scfd and the result of root cause
analysis must be reported in the semi-annual compliance reports.
Because affected flares are also affected fuel gas combustion devices, the root cause
analysis for SO2 emissions exceeding 500 lbs/day also applies to flares. However, compliance
with the 500 lb/day root cause analysis will also require continuous monitoring of total reduced
sulfur of all gases flared. Although all fuel gas combustion devices are required to comply with
continuous H2S monitoring of fuel gas, flares routinely accept gases from upsets, malfunctions
and startup and shutdown events, and H2S or sulfur monitoring is not specifically required for
these gases. In subpart Ja, we explicitly require TRS monitoring to ensure that the 500 lb/day
SO2 trigger is accurately measured. We also note that for affected flares, the RCA trigger is 500
lbs/day, not 500 lbs/day in excess of an emissions limit.
4-9
-------
4.2.6 Modification and Reconstruction Provisions
Existing affected facilities that commence modification or reconstruction after May 14,
2007, are subject to the final standards in 40 CFR part 60, subpart Ja. A modification is any
physical or operational change to an existing affected 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 affected 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.
In response to comments regarding the work practice standards for flares and fuel gas
producing units, we re-evaluated the work practice standards and have decided to define a flare
as the affected source rather than a fuel gas producing unit. The intermittent operation of a flare
makes it difficult to use the criteria of 40 CFR 60.14 to determine when a flare is modified;
therefore, we have specified in the final rule the criteria that define a modification to a flare. A
flare is considered to be modified if: (1) any piping from a refinery process unit or fuel gas
system is newly connected to the flare or (2) the flare is physically modified to increase flow
capacity.
Special provisions are included for NOx emissions from certain existing process heaters
that were modified or reconstructed between proposal and promulgation of this final rule in order
to avoid the retroactive application of the more stringent NOx emissions limit in the final rule.
Existing process heaters with a rated capacity of greater than 40 MMBtu/hr for which
modification or reconstruction commenced after May 14, 2007, and on or before the date of
publication of this final rule in the Federal Register are subject to the same NOx emission limit
as proposed (80 ppmv, dry basis corrected to 0% excess air on a 24-hour rolling average basis).
Petroleum refinery process units are subject to the final standards in 40 CFR part 60,
subpart Ja if they meet the criteria 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% 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).
4.3 Model Facility, Source Projections, and Cost Assumptions
EPA developed control options for this final to limit emissions of particulate matter,
sulfur dioxide, nitrous oxides and volatile organic compounds from new and modified processes
at petroleum refineries, including FCCUs, SRUs, cokers, and process heaters. EPA has estimated
4-10
-------
the costs of complying with the final NSPS under each of the options, using a model facility
approach. During the 5-year period of analysis (or the analysis of impacts in the fifth year after
proposal), EPA assumes the industry will invest in enough new processes to be equivalent to an
average of three new refineries per year, or a total of 15 new refineries. Further, the new sources
are assumed to be 40% new processes or facilities and 60% modified or reconstructed facilities
or processes.
Table 4-1 illustrates the assumptions used to characterize the model facilities and
processes upon which the costs of complying with the rule are based. More details follow on
these assumptions.
To project the number of new, reconstructed, and modified process units over the next 5
years, we used many of the same assumptions as in the analysis for subpart GGG, including the
average number of process units at a refinery (Parrish, Randall, and Coburn, 2006). That analysis
assumes that there are 0.8 FCCUs and 0.4 coking units per refinery and that 15 refineries' worth
of process units become subject (i.e., are either new, reconstructed, or modified) over the 5 years
following proposal. We also assumed that 40% of the FCCUs are new and 60% are reconstructed
or modified.
4.3.1 FCCUs
We identified four scenarios to characterize FCCUs, one for new process units and three
for reconstructed and modified process units. In this analysis, currently refers to the situation
prior to new NSPS requirements and baseline refers to the requirements if no new standards are
implemented(in most cases, baseline is compliance with subpart J).
1. New (baseline = comply with subpart J)
2. Currently subject to subpart J (baseline = continue to comply with subpart J)
3. Currently subject to consent decree (baseline = continue to comply with consent
decree requirements, assuming these are equal to or more stringent than subpart J)
4. Currently subject to MACT (baseline = comply with subpart J)
We assumed that the currently existing process units that are reconstructed or modified are
broken down into Scenarios 2, 3, and 4 as follows:
\
¦ 10% are subject to MACT and not subpart J (Scenario #4).
4-11
-------
Table 4-1. Model Facility Descriptions
Number
Model Facility Description Items Included in the Costs Facilities
New facility/major new expansion
¦ FCCU PM, S02, and NO, controls
¦ 11 new process heaters with NOx controls
¦ New fuel gas combustion device
¦ Implement work practice standards3
2
New FCCU
¦ FCCU PM, S02, and NO„ controls
¦ 2 new process heaters with NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
1
New processes
¦ 2 process heaters with NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
6
Small business with new processes
¦ SRU tail gas treatment
¦ 3 process heaters with retrofit NO„ controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
1
Multi-process and FCCU revamp
¦ FCCU PM, S02, and NOx retrofit controls
¦ 9 new process heaters with NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards8
3
Multi-process and FCCU revamp (no
¦ FCCU PM, S02, and NOx controls
2
delayed coker unit [DCU])
¦ 9 new process heaters with NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
FCCU revamp
¦ FCCU PM, S02, and NOx controls
¦ 2 process heaters with retrofit NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
4
Fluid coking unit revamp
¦ FCCU PM, S02, and NOx controls
¦ 7 new process heaters with NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
1
Modified processes
¦ 3 process heaters with retrofit NOx controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
8
Small business with modified processes
¦ SRU tail gas treatment
¦ 3 process heaters with retrofit NO, controls
¦ Modified/reconstructed fuel gas combustion device
¦ Implement work practice standards6
1
a Work practice standards include: fuel gas recovery for fuel gas producing units; flare minimization plan for
planned start-up and shutdown; sulfur shedding plan; root-cause analysis; and delayed coking depressurization gas
recovery.
b All of the work practices above except delayed coking depressurization gas recovery.
4-12
-------
¦ 76.5% are units subject to a consent decree (Scenario #3). For purposes of this
analysis, average consent decree requirements are assumed to be 1.0 kg PM/Mg coke
burn (using Method 5B or 5F), 50 ppmv SO2 over a 7-day average, and 25 ppmv SO2
over a 365-day average.
¦ The remaining 13.5% are subject to subpart J (Scenario #2).
The assumptions outlined above translate into the following values:
¦ 12 total new, reconstructed, or modified FCCUs (multiply estimate of 0.8 FCCU per
refinery by 15 refineries' worth of process units)
¦ 4.8 FCCUs are new
- ¦ 7.2 FCCUs are reconstructed or modified
- 1 is currently subject to subpart J
- 5.5 are currently subject to consent decree
- 0.7 are currently subject to MACT only
4.3.2 Fluid Coking Units (FCU)
We assumed that there will be six new, reconstructed, or modified coking units over the
next 5 years. Based on industry trends, we anticipate that five of these will be delayed coking
units and only one will be a fluid coking unit.1 We assumed that the single fluid coking unit will
become subject through modification or reconstruction rather than new construction. At baseline,
this fluid coking unit would comply with subpart J, which includes no requirements for coking
units.
We assumed that 30 refineries will construct, reconstruct, or modify a fuel gas
combustion device over the next 5 years. This estimate is based in part on the assumed average
numbers of process units at a refinery from the analysis for subpart GGG (Parrish, Randall, and
Coburn, 2006). Industry representatives have indicated that most fuel gas systems are
centralized. Therefore, the entire system that includes the new, reconstructed, or modified
combustion device will essentially have to meet subpart Ja standards in order for the new or
reconstructed fuel gas combustion device to comply with subpart Ja.
4.4 Emissions Estimation by Unit Type
For both FCCUs and fluid coking units, PM and SO2 controls were evaluated together
because a wet scrubber installed to reduce PM will also achieve SO2deductions. The other
control device considered for FCCUs was an electrostatic precipitator (ESP) for PM reduction
and catalyst additives for SO2 emission reduction. This option is not technically feasible for a
1 10-26 meeting minutes.
4-13
-------
fluid coking unit; therefore, the analysis for the one affected fluid coking unit assumed a wet
scrubber as the control device.
4.4.1 FCCUs
We assumed a model FCCU size of 50,000 barrels (bbl) per day. This model FCCU also
has a volumetric flow rate of 140,000 dry standard cubic feet per minute (dscfm) and a coke
burn-off rate of 800,000 pounds (lb) per day. It operates at 95% of capacity.
In order to determine emission reductions beyond subpart J for each option, we first
estimated emissions attributed to meeting subpart J. Based on industry trends and control device
capabilities, we assumed that 35% of the FCCUs would meet subpart J with a wet scrubber and
65% would meet subpart J using an ESP and catalyst additives. We assumed the basic model wet
scrubber could meet the subpart J PM limit with an 80% control efficiency and would average
about 25 ppmv SO2. The model ESP also had a control efficiency of 80% for PM. For SO2, we
calculated that the 9.8 kg/Mg coke burn is equivalent to about 265 ppmv.
We estimated emissions of PM as the total of filterable PM that is less than 10 .
micrometers (fim) in diameter (PM|0), filterable PM that is less than 2.5 (urn) in diameter
(PM2 5), and condensable PM. At baseline, an FCCU meeting subpart J with a wet scrubber
would emit 236 tons PM per year and an ESP would emit 305 tons PM per year. (A wet scrubber
has a lower operating temperature than an ESP, which provides improved removal of
condensable PM and results in lower PM emissions.) Based on the assumptions described above,
we estimated baseline PM emissions for the 12 FCCUs at 3,370 tons per year (1,350 tons per
year from new FCCUs and 2,020 tons per year from reconstructed and modified FCCUs). For
this model FCCU, we estimated emissions of SO2 as 1,540 tons per year for catalyst additives
meeting 265 ppmv and 145 tons per year for wet scrubbers and catalyst additives meeting 25
ppmv. Based on the assumptions described above, we estimated baseline SO2 emissions for the
12 FCCUs at 9,600 tons per year (5,050 tons per year from new FCCUs and 4,560 tons per year
from reconstructed and modified FCCUs).
To determine the emissions for each option, we assumed that the ratio of ESPs to wet
scrubbers chosen for new FCCUs would change depending on the particular emission limits
being considered. For example, as the SO2 limit tightens, a wet scrubber becomes more cost-
effective compared to the catalyst additives. On the other hand, we have no data to support an
assumption that wet scrubbers could achieve the Option 5 PM limit of 0.15 kg/Mg coke burn, so
we assumed that for Option 5, all FCCUs would be controlled with an ESP and catalyst
additives. In addition to these considerations, we considered for reconstructed and modified
4-14
-------
FCCUs with an existing control device whether cost-effectiveness or technical limitations for
each option would drive an operator to change the control device. For example, for Option 5, all
wet scrubbers would be removed in favor of ESPs that can meet the lower PM limit.
For each of the five options, we again estimated PM emissions as a total of filterable
PM 10 and PM25 and condensable PM. The total values vary for each option and for the specific
control device chosen. In addition to the SO2 emissions described for baseline, we estimated
emissions of 290 tons per year for catalyst additives meeting 50 ppmv. We also assumed that wet
scrubbers designed to meet 0.5 kg PM/Mg coke burn would achieve 12.5 ppmv SO2, which we
calculated to be equivalent to 73 tons per year.
4.4.2 FCUs
We assumed a model fluid coking unit size of40,000 bbl/day. This model coking unit
also has a volumetric flow rate of 200,000 dscfm. At baseline, there are no requirements for fluid
coking units, so there are no emission reductions for either PM or SO2. For Option 1, we
assumed that a basic wet scrubber would be chosen. We estimated emission reductions of 1,710
tons PM per year (based on 84% efficiency) and 20,600 tons SO2 per year (based on estimates of
94% efficiency and uncontrolled SO2 emissions of 3.0 lb/bbl).2 For Option 2, we assumed that an
enhanced wet scrubber would be chosen to meet the emission limits. We estimated emission
reductions of 1,970 tons PM per year (based on 97% efficiency) and 21,200 tons SO2 per year
(also based on an estimate of 97% efficiency).
We assumed that an average amine treatment system for fuel gas combustion devices
would have an average gas flow rate of 10,000 standard cubic feet per minute (scfm). We
developed this model system based on information from various sources (Polasek, Bullin, and
Donnelly, 1982; Fedich, Woerner, and Chitnis, 2004; Voltz, Corley, and Fedich, 2004). Based on
this flow rate and an emission limit of 20 ppmv, one system would emit 0.27 pounds (lbs) of SO2
per minute (min), or 70 tons per year (tons/yr). Nationwide (i.e., for 30 systems), the total
emissions are 2,100 tons per year. The emissions from one system, nationwide emissions, and
the reduction from the baseline for each of the three options are shown in Table 4-2.
2 Valero submittal (test report specifically).
4-15
-------
Table 4-2. Emissions for Fuel Gas Combustion Devices
Emissions from One System
Nationwide Emissions
(tons S02/yr)
Reduction from Baseline
(tons S02/yr)
Option (lb S02/min) (tons S02/yr)
2
3
0.13
, 0.10
0.07
35
26
17
1,050
786
524
1,050
1,310
1,570
4.5 Control Technologies in Analysis
The following NOx control techniques were included in the cost analysis for the control
¦ Flue Gas Recirculation. Flue gas recirculation (FGR) uses flue gas as an inert
material to reduce flame temperatures. In a typical flue gas recirculation 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 oxygen content
of the inlet air stream to the burner. The lower oxygen level in the combustion zone
reduces flame temperatures, which in turn reduces NOx emissions. The normal NOx
control efficiency range for FGR is 30% to 50%. When coupled with low-NOx
burners (LNB) the control efficiency increases to 50%-72%.
¦ Low-NOx Burners. Low-NOx burner (LNB) technology utilizes advanced burner
design to reduce NOx formation through the restriction of oxygen, flame temperature,
and/or residence time. The two general types of low NOx burners are staged fuel and
staged air burners. Staged fuel LNBs are particularly well-suited for boilers and
process heaters burning process and natural gas, which generate higher thermal NOx.
The estimated NOx control efficiency for LNBs where applied to petroleum refining
fuel burning equipment is generally around 40%.
¦ Ultra-low NOx Burners. Ultra-low NOx burners (ULNBs) may incorporate a variety
of techniques including induced flue gas recirculation (IFGR), steam injection, or a
combination of techniques. These burners combine the benefits of flue gas
recirculation and low-NOx burner control technologies. Rather than a system of fans
and blowers (like FGR), the burner is designed to recirculate hot, oxygen-depleted
flue gas from the flame or firebox back into the combustion zone. This leads to a
reduction in the average oxygen concentration in the flame without reducing the
flame temperature below temperatures necessary for optimal combustion efficiency.
The estimated NOx control efficiency for ULNBs in high temperature applications is
50%. Newer designs have yielded efficiencies of between 75% and 85%. When
coupled with selective catalytic reduction, efficiencies in the range of 85% to 97%
can be obtained. t
¦ Controlling Excess Oxygen in Complete Combustion FCCU Catalyst
Regenerators. Most of the previous control options are specific to process heaters
and carbon monoxide (CO) boilers. However, controlling the oxygen concentration in
options:
4-16
-------
the FCCU regenerator exhaust at approximately 0.5% has been seen to reduce NOx
concentrations by 20% to 40% as compared to NOx concentrations when the
regenerator exhaust oxygen concentration is between 1% and 2%. As sucli, complete
combustion FCCU regenerators with active excess oxygen controls are expected to
have similar performance as partial combustion FCCUs followed by CO boilers that
use low-NOx burners or flue gas recirculation.
Selective Non-Catalytic Reduction. In the selective non-catalytic reduction (SNCR)
process, urea or ammonia-based chemicals are injected into the flue gas stream to
convert nitric oxide (NO) to nitrogen gas (N2) and water. Without the participation of
a catalyst, the reaction requires a high temperature range to obtain activation energy.
The optimum operating temperature for SNCR is 1,600°F to 2,100°F. At
temperatures above 2,000°F, NOx control efficiency decreases rapidly. The normal
NOx control efficiency range for SNCR is 50% to 70%. SNCR systems are usually
lower in capital cost than SCR systems for the same application. One advantage of
this technology is the fact that no liquid or solid waste is generated. SNCR
technology has been applied to CO boilers, process heaters and boilers in the
petroleum refining sector where control efficiencies are consistent with the range
mentioned above.
LoTOx™ Technology. The LoTOx™ process (i.e., low-temperature oxidation) is a
patented technology that uses ozone to oxidize NOx to nitric pentoxide and other
higher order nitrogen oxides, 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 NOx emission levels have been reduced to less than
20 parts per million by volume (ppmv), and often as low as 10 ppmv, when inlet NOx
concentrations ranged from 50 to 200 ppmv (an 80% to 90% reduction efficiency).
Selective Catalytic Reduction. Selective catalytic reduction (SCR) is a post-
combustion NOx control technology in which ammonia (NH3) is injected into the
post-combustion gas stream in the presence of a catalyst. A catalyst bed containing
metals in the platinum family is used to lower the activation energy required for NOx
decomposition. The reaction of NH3 and NOx is favored by the presence of excess
oxygen. The NH3 oxidation to NOx increases with increasing temperature. The
normal NOx control efficiency range for SCR is 70% to 90%. There are at least three
SCRs currently in-use at refineries to control FCCU NOx emissions.
Combination System. Combination systems have used combustion controls followed
by SCR or SNCR technology in order to reduce costs of NOx removal from a flue
gas. For example, LNB has been combined with SNCR technology to minimize the
capital and operating cost for NOx removal as well as improve the control efficiency.
Catalyst Additives. An additional NOx emission control option specific for the
FCCU is the use of catalyst additive, such as X-NOx and DENOX (from Grace-
Davison; Bruhin et al., 2003). Non-platinum combustion promoter additives appear to
achieve a 30% to 50% emission reduction. Additional catalyst additives have had
limited success at further reducing the NOx emissions from the FCCU, and the results
of these other additives have been quite varied.
4-17
-------
Not all of these control technologies are applicable to all units. For example, use of catalyst
additives is applicable only to FCCUs, while the use of low-NOx burners are not applicable to
complete combustion FCCU catalyst regenerators. For each source type, costs for four control
scenarios were developed in 2005 dollars; the control scenarios ranged between 35% and 95%
NOx emission reduction efficiencies. As the baseline emissions of different FCCUs can span a
fairly significant range of outlet NOx concentrations, representative baseline concentrations were
assigned a weighting factor to simulate the distribution of baseline NOx emissions. For each
representative baseline NOx concentration, the control scenario needed to achieve a given
emission limit was assigned to the fraction of FCCUs represented by that concentration. The
overall costs for a given scenario were then calculated based on the weighting attributed to that
uncontrolled NOx concentration range. Although this basic approach was used for each NOx
emission source type, the specific costing methodologies for each of the three types of sources
are presented in separate sections to clearly identify the differences in the costs developed for the
different sources.
For SO2 and VOC, and PM control technologies, please refer to the technical memoranda in the
public docket for this rulemaking.
4.6 Cost Analysis for Control Options
EPA examined control options for individual refinery processes as well as work practice
options affecting the refinery as a whole. This section describes these control options and the
estimated costs of implementing them.
The costs presented in this section are calculated based on the control cost methodology
presented in the EPA (2002) Air Pollution Control Cost Manual prepared by the U.S.
Environmental Protection Agency.3 This methodology sets out a procedure by which capital and
annualized costs are defined and estimated, and this procedure is often used to'estimate the costs
of rulemakings such as this one. The capital costs presented in this section are annualized using a
7% interest rate, a rate that is consistent with the guidance provided in the Office of Management
and Budget's (OMB's) (2003) Circular A-4.4 Equipment lives for the control technologies
employed in this analysis can vary greatly (usually from 10 to 25 years).
3 Available on the Internet at http://eDa.iJOv/ttn/ciitc/nroducts.hlmNcccinfo.
4 Available on the Internet at hrtp://\v\v\v.whitehouse.gov/oinb/circulars/a004/a-4.ndf.
4-18
-------
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 over 80%
of the national refining capacity, are subject to consent decrees that limit the emissions from
subpart J process units.) In addition, we received a total of 46 comments during the public
comment periods associated with the proposed rule and NODA. These comments were received
from refineries, industry trade associations, and consultants; state and local environmental and
public health agencies; environmental groups; and members of the public. In response to these
public comments, most of the cost and emission reduction impact estimates were recalculated,
resulting in several changes to the final amendments and new standards. A summary of the
remainder of the comments received during the comment period and responses thereto can be
found in the docket for the final amendments and new standards (Docket ID No. EPA-OAR-
HQ-2007-0011). The docket also contains further details on all the analyses summarized in the
responses below.
Once we identified potential emission limjts for various process units, we evaluated each
limit in conjunction with control technology, costs, and emission reductions to determine the
Best Demonstrated Technology (BDT) for each process unit. In responding to the public
comments, we re-evaluated the costs and cost-effectiveness of the control options and re-
evaluated our BDT determinations. In our BDT determinations, we took all relevant factors into
account, including cost considerations, which are generally consistent with other Agency
decisions. It is important to note that, due to the different health effects associated with different
pollutants, the acceptable cost-effectiveness of a control option is pollutant dependent. These
pollutant-specific factors were considered in our BDT determinations.
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, as
mentioned earlier in this chapter, the standard metric for expressing costs and emission
reductions is the impact on all affected facilities in the fifth year after proposal. Details of the
calculations can be found in the public docket.
4-19
-------
4.7 Summary of Cost, Environmental, Energy, and Economic Impacts
We are presenting estimates of the impacts for the final 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 for flares and delayed coking units. 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 final subpart Ja and the current
1MSPS 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.
In order to determine the incremental costs and emission reductions of this final 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 final subpart Ja. Sources
that are modified or reconstructed over the next 5 years must comply with subpart J in the
absence of final subpart Ja. We assumed that prior to reconstruction or modification, these
sources will either be subject to a consent decree (equivalent to more than 80% of the industry by
capacity), complying with subpart J or equivalent limits, and/or complying with 40 CFR part 63,
subpart UUU (MACT 11). 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 final subpart Ja were estimated from those starting points as well.
When considering and selecting emission limits for the final rule, we evaluated the cost-
effectiveness of each option for new sources separately from reconstructed and modified sources.
In most cases, our selections for each process unit and pollutant were consistent for modified and
reconstructed units and new units. In this section, we are presenting our costs and emission
reductions for the overall rule. We estimate that the final amendments for new and modified and
reconstructed sources together will reduce emissions of PM by 1,300 tons/yr, SO2 by 17,000
tons/yr, NOx by 11,000 tons/yr, and VOC by 1,400 tons/yr from the baseline. The estimated
increase in annual cost, including annualized capital costs, is $31,000,000 (2006 dollars). The
estimated nationwide 5-year incremental emissions reductions and cost impacts for the final
standards are summarized in Table 4-3. A summary of the impacts by all options considered in
the course of preparing the final NSPS, for new and for modified and reconstructed sources, is
available in Appendix B.
4-20
-------
4.7.1 Secondary Impacts
Indirect or secondary air quality impacts of this final rule will result from the increased
electricity usage associated with the operation of control devices. Assuming that plants will
purchase electricity from a power plant, we estimate that the final standards will increase
secondary emissions of criteria pollutants, including PM, SO2, NOx, and CO from power plants.
For new, modified or reconstructed sources, this final rule will increase secondary PM emissions
by 56 Megagrams per year (Mg/yr) (62 tons/yr); secondary SO2 emissions by about 1,400 Mg/yr
Table 4-3. National Incremental Emission Reductions and Cost Impacts for Petroleum
Refinery Units Subject to Final Standards Under 40 CFR Part 60, Subpart Ja
(Fifth Year After Proposal)*
Annual
Annual
Annual
Total
Total
Annual
Emission
Emission
Emission
Capital
Annual
Emission
Reductions
Reductions
Reductions
Cost-
Cost
Cost
Reductions
(tons
(tons
(tons
Effectiveness
Process Unit
($1,000)
($l,000/yr) (tonsPM/yr)
S02/yr)
NO,/yr)
VOC/yr)
(S/ton)
FCCU
8,500
6,400
240
4,300
2,600
890
FCU
14,000
4,000
1,000
5,900 1
660
530
SRP
1,700
730
420
1,700
Fuel gas
34,000
12,000
5,200
2,300
combustion
devices
Process heaters
23,000
12,000
7,500
1,600
Flaring
40,000
-6,600
80
5
206
-23,000
Delayed coking
17,000
1,600
440
25
3,400
units
Sulfur pits
8,300
1,000
300
*
3,400
Total
150,000
30,700
1,300
17,000
11,000
1,400
1,070
a All costs are in 2006 dollars.
(1,500 tons/yr); secondary NOx emissions by about 530 Mg/yr (580 tons/yr).
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 FCU 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 final rule will generate 1.6 billion gallons of water per year for
the 5 years following proposal. We also estimate that this final rule will generate 2,200 Mg/yr
(2,400 tons/yr) of solid waste over those 5 years.
4-21
-------
Energy impacts as defined in this preamble section consist of the electricity and steam
needed to operate control devices and other equipment that would be required under the final
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 final rule
will increase overall energy demand by about 410 gigawatt-hours per year (1,400 billion British
thermal units per year). An analysis of energy impacts that accounts for reactions in affected
markets to the costs of this final rule can be found in the section on Executive Order 13211 found
later in this R1A.
4.8 Limitations and Uncertainties Associated with the Cost Analyses
Limitations and uncertainties associated with the cost analyses presented above are
presented here.
1. Assumptions behind the cost savings estimated for this NSPS. These annualized cost
savings are driven by the recovery of natural gas associated with the minimization of
flaring. The calculation presumes that all refiners owning new and modified or
reconstructed that must comply with the NSPS will be able to accomplish this,
particularly since they are estimated to receive a return on their investment. However,
opportunity costs may exist to make such an. investment more financially difficult
than we may expect. Smaller refiners may have less capital available to them than
larger refiners, and may find the investment in necessary capital and development of a
flare gas minimization plan more challenging than larger refiners. In addition, there
may be varying abilities among refiners to develop such minimization plans with their
current labor force due to tightness in labor markets, though such difficulties could be
remedied beyond the short run. Finally, many refiners have a minimum rate of return
that must be met before an investment to expand capacity, or for other reasons, is
carried out. For example, the Ad Hoc Coalition of Small Business Refineries asserted
in a comment on the proposed petroleum refineries NSPS that an acceptable rate of
return of well above 10% served as a minimum for investment purposes at their
refineries.5 It also should be noted, however, that the natural gas price that underlies
the estimate of cost savings is $7/Mcf (1,000 cubic feet), which is the average natural
gas price in 2006, the base year for the cost analysis. Current spot prices for natural
gas exceed $10/Mcf, and there is potential for further increases in such natural gas
prices to as high as $20/Mcf in the future by some estimates.6 An increase in natural
gas price may lead to additional incentive for refiners to engage in flare gas
minimization.
2. Effect of consent decrees on overall impact estimates. As mentioned in the preamble
for this final NSPS, the estimates of impacts in this RIA are based on a baseline that
5 Comments by Ad Hoc Coalition of Small Business Refiners to U.S. EPA, August 27,2007. Comments on
Proposed subpart Ja Rule: Standards for New, Modified, and Reconstructed Process Units at Petroleum
Refineries. 72 FR 27178. p. 6.
6 Wall Street Journal, April 18, 2008. "Surge in Natural Gas Price Stoked by New Global Trade."
4-22
-------
includes consent decrees reached by the U.S. Department of Justice (DOJ) as part of
the Refinery Enforcement Initiative with refiners representing 77% of U.S. refinery
capacity. Consent decrees continue to take place, however. For example, Holly Oil
and Refining Co. in Woods Cross, UT, just entered into a consent decree this month
with the U.S. DOJ to reduce emissions of SO2, NOx, and VOC from their refinery.
With this act, refiners representing 87% of U.S refinery capacity are now under a
consent decree.7 As more refiners are covered by consent decrees, the impacts
associated with this NSPS will be less.
3. New source bias. An NSPS is an example of what is known as a "vintage
differentiated regulation (VDR)," which is a term often used to describe regulations
that are fixed with respect to the date of entry of regulated units, with later vintages
facing more stringent standards. Often, units produced before a given date are
exempted or "grandfathered" from regulation. Reasons that are given for their being
commonly applied in regulatory policy is that it is more efficient and cost-effective to
control a given amount of pollution at a new plant as compared to retrofitting an
existing plant. Also, there may be greater equity in not changing environmental
regulations for facilities that have already been built, and instead focus on only new
facilities. However, there may be incentive for existing firms to desire
implementation of VDRs, for these can be used to erect entry barriers to restrict
competition and protect rents created by existing command-and-control standards.
Thus, there are efficiency and equity considerations inherent in the setting of an
NSPS like this one for the petroleum refineries industry, as well as potential concerns
about it being a possible barrier to entry, though this depends on the degree of
stringency in the NSPS. This NSPS may create economic inefficiency by
discouraging technological innovation and investment/modernization among
regulated sources. Insofar as refiners that manage to avoid triggering NSPS will be
rewarded, while those that do invest/modernize in capacity expansions may trigger it,
the rule will impose efficiency losses on the economy that are not captured well by
the economic impact model used in the RIA All of these considerations could affect
the actual estimates of impacts associated with this NSPS; it may become more
difficult to build new and modified and reconstructed process units (cost savings
notwithstanding), while existing units and refiners may receive an advantage in terms
of some protection against entry of new firms and process units. Overall efficiency
from a social welfare standpoint may also be affected.
4. Projections of new and modified and reconstructed sources. We project that during
the 5-year period of analysis, EPA assumes the industry will invest in enough new
processes to be equivalent to an average of three new refineries per year, or a total of
15 new refineries. Further, the new sources are assumed to be 40% new processes or
facilities and 60% modified or reconstructed facilities or processes. Our baseline
estimates of emissions and process units, and thus our incremental analysis options
that are applied to these baseline estimates, depend strongly on these estimates. Many
of these estimates are taken from industry trade journals (e.g., Oil and Gas Journal).
1 Quoted from the U.S. EPA Web site for the Refinery Enforcement Initiative at
http://www.epa.gov/compliance/resources/cases/civil/caa/oil/index.html.
4-23
-------
The certainty of these estimates is only as high as that of the experts' capabilities that
prepare these estimates.
4.9 References
Bruhin et al., 2003, 17.
Fedich, R.B., A.C. Woerner, and G.K. Chitnis. May 2004. Selective H2S Removal. Hydrocarbon
Engineering, Vol. 9, No. 5, pp. 89-92.
Parrish, K., D. Randall, and J. Coburn. October 30, 2006. Data and Assumptions used in the
Equipment Leaks Cost Analysis for Petroleum Refineries. Memorandum to Karen
Rackley, EPA/SPPD. Docket Item No. EPA-HQ-OAR-2006-0699-0034.
Polasek, J.C., J.A. Bullin, and S.T. Donnelly. 1982. Alternative Flow Schemes to Reduce Capital
and Operating Costs of Amine Sweetening Units. Proceedings of the 1982AIChE Spring
National Meeting, New York, NY: American Institute of Chemical Engineers.
Stavins, Robert N. "The Effects of Vintage-Differentiated Environmental Regulation."
Resources for the Future, Discussion Paper 05-12, March 2005. Available at
.
Voltz, B.L., J.D. Corley, and R.B. Fedich. November 2004. Benefits of a TGCU Amine Solvent
Changeover. Sour Oil & Gas Advanced Technology 2004 International Conference.
4-24
-------
APPENDIX B:
SUMMARY OF SIGNIFICANT COMMENTS AND RESPONSES, AND RATIONALES
FOR NSPS EMISSION LIMITS
Below is a summary of the public comments and our responses to them organized by
source type and pollutant controlled. Within this appendix is the rationale for the emission limits
included in the final NSPS for each source type and pollutant.
B.l PM Limits for Fluid Catalytic Cracking Units
Comment: Several commenters opposed the proposed tightening of the FCCU PM
standards relative to subpart J and the concurrent change in PM monitoring methods. Some
commenters supported the co-proposal to keep the 1 lb/1,000 lb coke burn PM emission limit
based on Method 5B and/or 5F; other commenters either did not oppose or supported the 0.5
lb/1,000 lb coke burn emission limit for new "grassroots" units, provided EPA demonstrates it is
cost-effective and that the limit is based on EPA Method 5B or 5F (40 CFR part 60, Appendix
A-3).
Commenters stated that EPA should only impose the more stringent emission limits on
new construction because it is much more difficult and costly to meet the proposed emission
limits for modified or reconstructed equipment. Commenters suggested that if EPA does include
more stringent limits on modifications, it should exclude certain actions (like projects
implemented to meet consent decree requirements) from the definition of a modification.
Several commenters suggested that the costs in Table 11 of the proposal preamble are
significantly underestimated. Commenters contended that the single "model plant" approach
used in EPA's cost analysis does not realistically consider important factors such as the inherent
sulfur content of the feed, partial-burn versus full-burn regeneration, FCCU/regenerator size, and
sources that are already well-controlled due to other regulations. Commenters asserted that the
purchased equipment costs escalated from estimates that are 20 to 30 years old are
underestimated. Several commenters provided estimates of costs and emission reductions for
several actual projects, which they stated indicate that EPA's costs are significantly
underestimated and that the proposed standards are much less cost-effective than presented by
EPA.
A number of commenters asserted that the PM standards should be based on EPA
Methods 5B or 5F (40 CFR part 60, Appendix A-3), and not on EPA Method 5 of Appendix A-3
B-l
-------
to part 60. According to these commenters, the achievability of the proposed 0.5 lb/1,000 lb
coke burn PM limit based on EPA Method 5 is questionable because there are inadequate data on
FCCU using EPA Method 5, and controlling combined condensable and filterable PM to the 0.5
lb/1,000 lb coke burn level has not been demonstrated to be cost-effective.
On the other hand, several commenters stated that any PM limit must include
condensable and filterable PM as condensable PM account for a large portion of refinery PM
emissions and all condensable PM is PM that is less than 2.5 micrometers in diameter (PM25),
which has more adverse health impacts than larger particles; the commenters therefore agreed
with the use of EPA Method 5 to determine filterable PM and requested that EPA consider
Method 202 (40 CFR part 51, Appendix M) for condensable PM. Commenters also stated that
the limits for PM and SO2 in subpart Ja should apply to all new, reconstructed, and modified
FCCU. One commenter recommended that a total PM limit (filterable and condensable) be set at
1 lb/1,000 coke burn; another stated that the total PM limit, including both filterable and
condensable PM, should be 0.5 lb/1,000 lb coke burn, and EPA has not demonstrated that current
BDT cannot achieve this limit. Finally, one commenter suggested that EPA should evaluate the
cost of removing each pollutant (PM and SO2) separately.
Response: In response to these comments, we have revised our analysis to consider each
unique existing FCCU in the United States. By doing so, we fully account for plant size, partial-
burn versus full-burn regeneration, existing control configuration, and specific consent decree
requirements. (Details on the specific revisions to the analysis can be found in the docket.) With
a revised analysis, we were able.to more directly assess the impacts of process modifications or
reconstruction of existing equipment. We also assessed the effects of PM and SO2 standards
separately in this analysis.
In our revised analysis, we considered three options for PM: (1) maintain the existing
subpart J standard of 1.0 lb/1,000 lb of coke burn-off (filterable PM as measured by Method 5B
or 5F); (2) 0.5 lb/1,000 lb of coke burn-off (filterable PM as measured by Method 5B or 5F of
Appendix A-3 to part 60); and (3) 0.5 lb/1,000 lb of coke burn-off (filterable PM as measured by
Method 5 of Appendix A-3 to part 60). Similar to the analysis for the proposed standards, costs
and emission reductions for each option were estimated as the increment between complying
with subpart J and subpart Ja. We note that none of the available data suggest that a 0.5 lb/1,000
lb coke burn emission limit that includes both filterable and condensable PM as measured using
B-2
-------
EPA Method 202 is achievable in practice for the full range of facilities using BDT controls, so
we disagree with the comments suggesting this level is appropriate to consider as an option for a
total PM limit in this rulemaking.
Option 1 includes the same emissions and requirements for PM as the current 40 CFR
part 60, subpart J, so it will achieve no additional emissions reductions. The PM limit in Option
2 is the same numerical limit that was proposed in subpart Ja, but the PM emissions are
determined using Methods 5B and 5F (40 CFR part 60, Appendix A-3). These test methods are
commonly used for PM tests of FCCU and are the methods that were used to generate a majority
of the test data we reviewed. Option 3 is a limit of 0.5 lb/1,000 lb coke burn using Method 5 and
is the performance level that was proposed for subpart Ja.
The impacts of these three options for new FCCU are presented in Table B-l; the impacts
for modified and reconstructed FCCU are presented in Table B-2.
Table B-l. National Fifth Year Impacts of Options for PM Limits Considered for New
Fluid Catalytic Cracking Units Subject to 40 CFR Part 60, Subpart Ja"
Capital Cost
Total Annual Cost
Emission Reduction
Cost-EfTectiveness (S/ton)
Option
($1,000)
($l,000/yr)
(tons PM/yr)
Overall
Incremental
1
0
0
0
N/A
N/A
2
3,600
1,100
240
5,600
5,600
3
7,100
1,700
300
6,700
11,000
a PM cost-effectiveness calculated for PM-fine, assuming 83.3% of the PM is PM-fine.
Table B-2. National Fifth Year Impacts of Options for PM Limits Considered for
Reconstructed and Modified Fluid Catalytic Cracking Units Subject to 40
CFR Part 60, Subpart Jaa
Capital Cost
Total Annual Cost
Emission Reduction
Cost-Jiffectiveness (S/ton)
Option
($1,000)
($l,000/yr)
(tons PM/yr)
Overall
Incremental
1
0
0
0
N/A
N/A
2
75,000
12,000
690
21,000
21,000
3
100,000
15,000
810
23,000
37,000
0 PM cost-effectiveness calculated for PM-fine, assuming 83.3% of the PM is PM-fine.
The available data and impacts for the options considered suggest that BDT for new
FCCU is different than BDT for modified and reconstructed FCCU. For new FCCU, the costs
for Option 2 are reasonable compared to the emission reduction achieved. The incremental cost
between Option 2 and Option 3 of $11,000 per ton PM-fine would generally be considered
B-3
-------
reasonable, but there are uncertainties in the achievability of Option 3. The estimated PM
emission reduction achieved by Option 3 compared to Option 2 equals 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 (40 CFR part 60, Appendix A-3). Additionally, available test data indicate
that electrostatic precipitators (ESP) and wet scrubbers can reduce total filterable PM to
0.5 lb/1,000 lb of coke burn or less, as measured by Method 5-equivalent test methods.
Although there were few test data points using Method 5-equivalent test methods, we concluded
at proposal that both electrostatic precipitators and wet scrubbers can achieve this level of PM
emissions. However, the data supporting Option 3 are not extensive, and it is unclear at this time
whether a limit of 0.5 kg/Mg of coke burn as measured by Method 5 (40 CFR part 60, Appendix
A-3) could be met by all configurations of FCCU. In addition, while the Agency supports
reducing condensable PM emissions, the amount of condensable PM captured by Method 5 is
small relative to methods that specifically target condensable PM, such as Method 202 (40 CFR
part 51, Appendix M). We prefer to develop a single performance standard that considers all
condensable PM rather than implementing phased standards targeting different fractions of
condensable PM. Such an approach would be costly and inefficient. Therefore, we conclude
that Option 2, control of PM emissions (as measured by Methods 5B and 5F of Appendix A-3 to
part 60) to 0.5 lb/1,000 lb of coke burn or less, is BDT for newly constructed FCCU. This option
achieves PM emission reductions of 240 tons per year (tons/yr) from a baseline of 910 tons/yr at
a cost of $5,600 per ton of PM.
For modified and reconstructed FCCU, Option 1 is the baseline level of control
established by the existing requirements of subpart J. It will achieve no additional cost or
emission reduction. The overall costs and the incremental costs for Options 2 and 3 are
reasonable compared to the PM emission reduction; however, as with new FCCU, the
performance of Option 3 has not been demonstrated, so it is rejected. Most of the existing FCCU
that could become subject to subpart Ja through modification or reconstruction are either already
subject to subpart J or are covered by the consent decrees. The consent decrees are generally
based on the existing subpart J. Industry has made significant investments in complying with
these subpart J requirements which may be abandoned if they become subject to subpart Ja. In
addition, the additional costs could create a disincentive to modernize FCCU to make them more
energy efficient or to produce more refined products. For these reasons, we reject Option 2 for
B-4
-------
modified or reconstructed FCCU and conclude that control of PM emissions (as measured by
Methods 5B and 5F of Appendix A-3 to part 60) is 1.0 lb/1.000 lb of coke burn or less is BDT
for reconstructed and modified FCCU.
B.2 SO2 Limits for Fluid Catalytic Cracking Units
Comment: Several commenters supported the co-proposal for modified and
reconstructed FCCU to meet subpart J and not the 25 ppmv 365-day rolling average limit for
SO2. Commenters provided data to suggest that the retrofits of existing sources are not cost
effective, particularly if catalyst additives cannot be used. The current subpart J includes three
compliance options: (1) if using an add-on control device, reduce SO2 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 SO2) 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. Several commenters objected to the elimination of the additional compliance options
in the existing subpart J for subpart Ja because: (1) there are no data to show that the SO2 limits
proposed in subpart Ja are BDT for all FCCU regenerator configurations; (2) the three options
are already established as BDT and, therefore, the CAA requires that EPA make them available;
and (3) the substantial cost and other burdens for a reconstructed or modified FCCU already
complying with one of the alternative options in subpart J to change to daily monitoring by
Method 8 (40 CFR part 60, Appendix A-4) or to install CEMS were not addressed in the
proposal.
One commenter supported the proposed SO2 limit under Ja for new "grassroots" FCCU if
the standard is demonstrated to be cost-effective.
Response: As acknowledged in the previous response on PM standards for FCCU, we
completely revised our impacts analysis to evaluate SO2 standards for every existing FCCU that
may become subject to subpart Ja through modification or reconstruction. We did not have
access to the inherent sulfur content of the feed for each FCCU so SO2 emissions are still
estimated using average emission factors relevant to the type of control device used for FCCU
not subject to consent decree requirements. Nonetheless, we significantly revised the impact
analysis to fully account for FCCU-specific throughput, existing controls, and consent decree
requirements. (Details oh the specific revisions to the analysis can be found in Docket ID No.
EPA-HQ-OAR-2007-0011.) We evaluated two options: (1) current subpart J, including all three
B-5
-------
compliance options; and (2) 50 ppmv SO2 on a 7-day average and 25 ppmv on a 365-day
average. Data are not available on which to base a more stringent control level.
Option 1 includes the same emissions and requirements as the current 40 CFR part 60,
subpart J, so it will achieve no additional emissions reductions. Based on information provided
by vendors and data submitted by petroleum refiners, Option 2 can be met with catalyst additives
or a wet scrubber. Of 38 FCCU currently subject to a 50/25 ppmv SO2 limit through consent
decrees, 26 used wet scrubbers and 12 used catalyst additives or other (unspecified) techniques.
Given the number of FCCU currently meeting the 50/25 ppmv SO2 emission limit, we conclude
that this limit is technically feasible.
The data in the record suggest that all systems with wet scrubbers can meet the 50/25
ppmv SO2 emission limit with no additional cost. Further, based on information from the
consent decrees, we believe that the owner or operator of an existing FCCU that does not already
have a wet scrubber and is modified or reconstructed such that it becomes subject to subpart Ja
can use catalyst additives to meet the 50/25 ppmv SO2 emission limit. Therefore, the cost of
Option 2 is calculated using catalyst additives as the method facilities choose for meeting the
standard. We reject the idea tha't the 90 percent control efficiency, the 9.8 kg/Mg coke burn-off
limit, or the 0.3 weight percent sulfur content alternatives are equivalent to the 50/25 ppmv SO2
emission limit. Based on the original background document for the subpart J standards, these
alternatives are expected to have outlet SO2 concentrations of 200 to 400 ppmv. In reality the
currently used wet scrubbers and catalyst additives achieve much higher SO2 removal
efficiencies and much lower outlet SO2 concentrations. The impacts of these options are
presented in Table B-3.
Table B-3. National Fifth Year Impacts of Options for $02 Limits Considered for New,
Reconstructed, and Modified Fluid Catalytic Cracking Units Subject to 40
CFR Part 60, Subpart Ja
Capital Cost
Total Annual Cost
Emission Reduction
Cost-Effectiveness ($/ton)
Option
(SI,000)
($l,000/yr)
(tons S02/yr)
Overall Incremental
1
0
0
0 '
N/A N/A
2
0
3,000
4,400
700' 700
Based on the data we reviewed to select the options and the estimated impacts of those
options, we conclude that Option 2, control of SO2 emissions to 25 ppmv or less averaged over
365 days and 50 ppmv or less averaged over 7 days, is technically feasible and cost-effective for
B-6
-------
new, reconstructed, and modified fluid catalytic cracking units. This option has no capita! cost
and achieves SO2 emission reductions of 4,400 tons/yr from a baseline of 5,900 tons/yr at a cost
of $700 per ton of SO2. Therefore, we conclude that control.of SO2 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.
B.3 NO, Limit for Fluid Catalytic Cracking Units
Comment: Several commenters stated that they would support a NOx limit of 80 ppmv
for new sources only, provided a corrected impact analysis considers the different characteristics
of FCCU and demonstrates that the NOx limit for new sources is truly cost-effective.
Commenters supported the co-proposal for modified and reconstructed FCCU to meet subpart J
and not be subject to a NOx emission limit. A few commenters provided cost data showing the
cost of NOx controls is high for modified and reconstructed units due to the high cost and space
needed for add-on controls. The commenters also stated that a large number of existing FCCU
in the U.S. are covered by consent decrees, so significant NOx reductions have already been (or
will soon be) achieved, and an additional incremental reduction to 20 or 40 ppmv over a 365-day
average are not widely demonstrated and would not be cost-effective.
One commenter stated that selective noncatalytic reduction (SNCR), selective catalytic
reduction (SCR), and catalyst additives have not been demonstrated over significant periods of
operational life. Commenters also cited environmental side-effects, such as the generation of
ammonia compounds that contribute to condensable PM emissions, as a reason not to require
these types of controls. Commenters also asserted that technologies like flue gas recirculation or
advanced burner design are typically only cost-effective for new units and may be technically
infeasible for existing FCCU.
One commenter suggested that if a limit is necessary for modified or reconstructed
FCCU, recent catalyst additive trials support an emission limit of approximately 150 ppmv on a
7-day rolling average; this limit would only be achievable if a 24-hour CO averaging time was
provided since lowering NOx tends to increase CO emissions in FCCU. The commenter noted
that this limit is equivalent to the 0.15 pounds per million-British thermal units (Ib/MMBtu)
standard for reconstructed and modified heaters and boilers in NSPS subpart Db.
Other commenters supported the inclusion of a NOx limit for FCCU and opposed the co-
proposal of no NOx standard for modified and reconstructed FCCU. These commenters also
B-7
-------
recommended more stringent NOx limits for FCCU and stated that 80 ppmv does not represent
an adequate level of control given the evolution of emerging technologies. In addition, a BDT of
80 ppmv on 7-day rolling average does not look "toward what may be fairly projected for the
regulated future" as required by Portland Cement I (486 F. 2d 375 at 384 (D.C. Cir. 1973)) and
other court decisions. The commenters disagreed with the feasibility and cost analyses for
modified and reconstructed FCCU and stated that FCCU under a consent decree are achieving
lower levels than the 80 ppmv proposed by EPA. Given the significant hazards to human health
and the environment posed by NOx emissions, the commenters recommended limits of 20 ppmv
over a 365-day rolling average and 40 ppmv over a 7-day rolling average for all FCCU. The
commenters noted that these limits have been successfully achieved under consent decrees and
they are technically feasible on new units at reasonable costs without additional controls.
Response: As shown by the disparate comments received, many commenters suggest
lower NOx emission limits are achievable, while other commenters do not believe the proposed
NOx emission limits are cost-effective. While we do acknowledge that lower NOx emission
limits are technically achievable, the incremental cost of achieving these lower limits was high
when we evaluated options for the proposed standards. Therefore, we concluded at proposal that
20 or 40 ppmv NOx limits were not BDT. In our BDT assessment, we evaluated the various
methods to meet alternative NOx limits as BDT rather than identifying one technology. One of
the reasons for this is that each technology has its own advantages and limitations. While non-
platinum oxidation promoters and advanced oxidation controls do not achieve the same reduction
in NOx emissions as add-on control devices such as SCR, they do so without any significant
secondary impacts. The added NOx reduction of SCR and SNCR must be balanced with these
secondary impacts. Part of the basis for selecting control methods to achieve an 80 ppmv NOx
emission limit as BDT included both cost and secondary impacts. This approach is necessary
when conducting our BDT analysis, thus ensuring the best overall environmental benefit from
the subpart Ja standards.
To ensure that we addressed the commenters' concerns, we re-evaluated the impacts for
FCCU NOx controls. We also collected additional data from continuous NOx monitoring
systems for a variety of FCCU NOx control systems. These data suggest that as refiners gain
more experience with the NOx control systems (including catalyst additive improvements), NOx
control performance has improved over the past year or two. These data suggest that the
B-8
-------
achievable level for combustion controls and catalyst additives is 80 ppmv and the achievable
level for add-on control systems is 20 ppmv. Therefore, we evaluated three outlet NOx emission
level options as part of the BDT determination: (1) 150 ppmv; (2) 80 ppmv; and (3) 20 ppmv.
Each NOx concentration is averaged over 7 days. To estimate impacts for Option 1, we
estimated that some units have current NOx emissions below 150 ppmv. and all other units can
meet this level with combustion controls such as limiting excess O2 or using non-platinum
catalyst combustion promoters and other NOx-reducing catalyst additives in a complete
combustion catalyst regenerator or a combination of NOx-reducing combustion promoters and
catalyst additives with low-NOx burners (LNB) in a CO boiler after a partial combustion catalyst
regenerator. Data collected from FCCU complying with consent decrees show that Option 2 can
also be met using combustion controls; therefore, we estimated impacts for Option 2 using a
similar method as Option 1. The main difference is that a larger number of FCCU must use
combustion controls to meet the emission limit (i.e.. the FCCU with current NOx emissions
between 150 and 80 ppmv would not need controls under Option I but would need controls
under Option 2). Option 3 is the level at which we expect all units to install more costly control
technology such as LoTOx™ or SCR.
The estimated fifth-year emission reductions and costs for each option for new FCCU are
summarized in Table B-4; the impacts for modified and reconstructed FCCU are summarized in
Table B-5.
Table B-4. National Fifth Year Impacts of Options for NOx Limits Considered for New
Fluid Catalytic Cracking Units Subject to 40 CFR Part 60, Subpart Ja
Option
Capital Cost
(SI,000)
Total Annual Cost
($l,000/yr)
Emission Reduction
(tons NO,/yr)
Cost-Effectiveness ($/ton)
Overall Incremental
1
860
320
370
880
880
2
1,200
640
860
750
650
3
12,000
3,600
1,400
2,600
5,800
B-9
-------
Table B-5. National Fifth Year Impacts of Options for NOx Limits Considered for
Modified and Reconstructed Fluid Catalytic Cracking Units Subject to 40
CFR Part 60, Subpart Ja
Capital Cost
Total Annual Cost
Emission Reduction
Cost-Effectiveness ($/ton)
Option
($1,000)
($l,000/yr)
(tons NO,/yr)
Overall
Incremental
i
2,800
1,000
860 v
1,200
1,200
2
3,700
1,600
1,800
920
660
3
45,000
11,000
3,200
3;600
6,800
Options 1 and 2 provide cost-effective NOx control with limited or no secondary impacts.
The costs of Option 1 and Option 2 are commensurate with the emission reductions for new
FCCU as well as modified and reconstructed FCCU. Option 3 would impose compliance costs
that are not warranted for the emissions reductions that would be achieved, as shown by the
incremental cost-effectiveness values of about $6,000 per ton of NOx emission reduction
between Option 2 and Option 3.
I
In evaluating these options, we also considered the secondary impacts. In addition to the
direct PM impacts of SNCR and SCR, SCR and LoTOx™ units require additional electrical
consumption. The increased energy consumption for Option 3 is 40,000 MW-hr/yr for new,
modified, and reconstructed units. We also evaluated the secondary PM, SO2, and NOx emission
impacts of the additional electrical consumption for Option 3. Based on the energy impacts,
Option 3 will generate secondary emissions of PM, SO2, and NOx of 6, 150, and 57 tons/yr,
respectively.
Based on the impacts shown in Table B-4 and Table B-5, and taking secondary impacts
into account, we conclude that BDT is Option 2, a NOx emission limit of 80 ppmv, for all
affected FCCU. For new FCCU, this option achieves NOx emission reductions of 860 tons/yr
from a baseline of 1,500 tons/yr at a cost of $750 per ton of NOx. For modified and reconstructed
FCCU, this option achieves NOx emission reductions of 1,800 tons/yr from a baseline of 3,600
tons/yr at a cost of $920 per ton ofNOx.
B.4 PM and SO2 Limits for Fluid Coking Units
Comment: Several commenters stated that EPA's proposed standards for FCU under
subpart Ja are inappropriate and not cost-effective. Commenters asserted that based on the
significant differences between FCU and FCCU operations, a separate BDT determination is
needed for FCCU and FCU. Commenters stated that an FCU has higher particulate loading; a
B-10
-------
heavier feedstock that typically contains a higher concentration of sulfur, increasing the SO2 and
sulfur trioxide (SO3) emissions: and a wider range of feedstocks with considerable variability in
the nitrogen content.
The commenters noted that the impacts analysis performed for the FCU has shortcomings
similar to those in the impacts analysis for FCCU (e.g., the analysis did not properly consider the
additional costs and technical difficulties of meeting the proposed emission limits for modified
or reconstructed sources, existing units are already controlled and thus the emission reductions
have already been achieved). One commenter provided site-specific engineering cost estimates
to indicate that the PM controls are much less cost-effective than EPA estimates. The
commenter requested that EPA consider instances when wastewater limitations require
regenerative wet scrubbers and amend the impact estimates accordingly. One commenter stated
that a newly installed regenerative wet scrubber system on an existing FCU could not meet the
proposed Ja PM standards.
Response: As described in the preamble to the proposed standards, the original analysis
assumed that one of the larger existing FCU will become a modified or reconstructed source in
the next 5 years. However, the two larger FCU in the U.S. are both subject to consent decrees:
one has installed controls and the other is in the process of installing controls. The remaining
two FCU are significantly smaller than the original model FCU; therefore, a new analysis was
conducted using a smaller model FCU indicative of the size of the two remaining FCU that are
not subject to consent decree requirements. In our new analysis, this FCU has approximately
one-half the sulfur content as the larger FCU for which we have data, based on information
received regarding the variability in sulfur content across different FCU in the public comments.
In addition to revising our impact analysis, we also collected additional source test data
from the one FCU operating a newly installed wet scrubber system to better characterize the
control system's performance. At proposal, we had one FCU source test from this source, which
suggested that the FCU wet scrubber could meet a PM limit of 0.5 lb/1,000 coke burn. However,
following proposal, we received an additional performance test for this same FCU wet scrubber
with an emission rate between 0.5 and 1.0 lb/1,000 lb coke burn. There was no indication of
unusual performance during either of these two tests, so we conclude that these tests demonstrate
the variability of the emission source and control system. Based on the available data, therefore,
we conclude that an appropriate PM performance level to consider for a BDT analysis is 1.0
B-ll
-------
lb/1,000 lb coke burn using EPA Method 5B (40 CFR part 60, Appendix A-3) for a FCU with a
wet scrubber. We also conclude that the PM emission limit initially proposed for FCU had not
been adequately demonstrated as an emission limit with which one must comply at all times.
Using our revised model FCU and based on the additional source test data, we re-
evaluated BDT for PM and SO2 emissions from FCU based on two options: (1) no new
standards, or current subpart J; and (2) a PM limit of 1.0 lb/1,000 lb coke burn (as measured
using Methods 5B and 5F of 40 CFR part 60, Appendix A-3), a short-term SO2 limit of 50 ppmv
averaged over 7 days, and a long-term SO2 limit of 25 ppmv averaged over 365 days. Unlike the
FCCU, catalyst additives cannot be used in a FCU to reduce SO2, so a wet scrubber is the most
likely technology (and the one demonstrated technology) that would be used to meet the PM and
SO2 limits of Option 2. Therefore, we estimated costs for an enhanced wet scrubber to meet both
the PM and SO2 limits. The resulting emission reductions and costs for both of the options are
shown in Table B-6.
Table B-6. National Fifth Year Impacts of Options for PM and SO2 Limits Considered for
Fluid Coking Units Subject to 40 CFR Part 60, Subpart Ja
Emission
Emission
Cost-Effectiveness
Capital Cost
Total Annual
Reduction (tons
Reduction (tons
(S/ton PM and
Option
(SI,000)
Cost ($l,000/yr)
PM/yr)
S02/yr)
so2)
1
0
0
0
0
N/A
2
10,000
3,200
1,000
5,900
460
2a
100,000
18,600
1,000
5,900
2,700
One commenter indicated that we should consider the costs of a regenerative wet
scrubber. This type of system is not needed in most applications, however, in the event such a
system were needed, we estimated the cost of a regenerative wet scrubber to meet Option 2. The
results of this analysis are also provided in Table B-6 as Option 2a. As seen in Table B-6, even
under the most conservative assumptions the costs associated with the PM and SO2 emission
reductions are reasonable.
Based on the available technology and the costs presented in Table B-6, we conclude that
BDT is Option 2, which requires technology that reduces PM emissions to 1.0 kg/Mg of coke
burn and reduces SO2 emissions to 50 ppmv averaged over 7 days and 25 ppmv averaged over
365 days. This option achieves PM emission reductions of 1,000 tons/yr from a baseline of 1,100
tons/yr and SO2 emission reductions of 5,900 tons/yr from a baseline of 6,100 tons/yr at a cost of
$460 per ton of PM and SO2 combined.
B-12
-------
B.5 NOx Limit for Fluid Coking Units
Comment: A number of commenters opposed the co-proposal of no NOx standard for
FCU. and some disagreed with EPA's 80 ppmv NOx limit for FCU. These commenters
recommended limits of 20 ppmv as a 365-day rolling average and 40 ppmv as a 7-day rolling
average for FCU, as has been successfully achieved under consent decrees. The commenters
noted that these limits are achievable on new units without additional controls. r
One commenter supported the co-proposal that no new NOx standard be established for
FCU.
Response: Similar to the revised analysis for PM and SO2 impacts, we re-evaluated BDT
for the FCU NOx controls for a smaller modified or reconstructed FCU. We evaluated three
options: (1) no new standards, which is the current subpart J; (2) outlet NOx concentration of 80
ppmv; and (3) outlet NOx concentration of 20 ppmv. Similar to the analysis for FCCU NOx and
depending on the baseline emissions for the FCU, we anticipate that Option 2 can be met using
combustion controls and Option 3 will require add-on control technology. The results of this
analysis are shown in Table B-7.
Table B-7. National Fifth Year Impacts of Options for NO* Limits Considered for Fluid
Coking Units Subject to 40 CFR Part 60, Subpart Ja
Capital Cost
(SI ,000)
Total Annual Cost
($l,000/yr)
Emission
Cost-Effectiveness ($/ton)
Option
Reduction
(tons NO, /yr)
Overall
Incremental
1
0
0
0
N/A
N/A
2
3,700
850
660
1,300
1,300
3
6,000
1,300
750
1,700
5,000
The costs for Option 1 and Option 2 are commensurate with the emission reductions, but
the incremental impacts for Option 3 are not reasonable, as shown in Table B-7. Option 3
achieves an additional 90 tons per year NOx reduction, but the incremental costs between options
2 and 3 of achieving this reduction is $5,000 per ton of NOx removed. The cost of achieving this
12 percent additional emission reduction nearly triples the total annualized cost of operating the
controls. As with FCCU, the add-on NOx controls for FCU have increased energy requirements
and secondary air pollution impacts. Based on these projected impacts, we support our original
determination that BDT is Option 2, or technology needed to meet an outlet NOx concentration
of 80 ppmv or less. This option achieves NOx emission reductions of 660 tons/yr from a baseline
of 800 tons/yr at a cost of $1,300 per ton of NOx.
B-13
-------
B.6 SO2 Limit for Small Sulfur Recovery Plants (SRP)
Comment: One commenter stated that no new requirements should be added for SRP
less than 20 LTD (small SRP) because the controls are not cost-effective. The commenter
provided data on tail gas treatment projects but noted that these costs are for large SRP, and
controls for small SRP will be less cost-effective. Several commenters noted that if EPA does
establish standards for small SRP, the monitoring and compliance evaluation methods for the 99
percent control standard are not clearly specified in the rule and could create difficulties in
documenting compliance for small Claus plants. Therefore, the small SRP should be allowed to
comply with the 250 ppmv SO2 emission limit provided to large SRP. One commenter
suggested that non-Claus units should be subject to a 95 percent recovery efficiency standard.
Response: To ensure that we addressed the commenters' concerns regarding cost-
effectiveness, we re-evaluated the impacts for small SRP. We adjusted our cost estimates
upward based on capital costs provided by industry representatives. We evaluated three SO2
control options as part of the BDT determination for small SRP: (1) no new standards, or current
subpart J; (2) 99 percent sulfur recovery; and (3) 99.9 percent sulfur recovery. As noted in the
preamble to the proposed standards (section V.D), the 99 percent and 99.9 percent recovery
levels are achievable for SRP of all sizes by various types of SRP or tail gas treatments.
The estimated fifth-year emission reductions and costs for new SRP are summarized in
Table B-8; the impacts for modified and reconstructed SRP are summarized in Table B-9. These
values reflect the impacts only for SRP smaller than 20 LTD; there are no additional cost impacts
for larger Claus units because they would already have to comply with the existing standards in
subpart J.
Table B-8. National Fifth Year Impacts of Options for SO2 Limits Considered for New
Small Sulfur Recovery Plants Subject to 40 CFR Part 60, Subpart Ja
Option
Capital Cost
($1,000)
Total Annual Cost
($l,000/yr)
Emission Reduction
(tons S02/yr)
Cost-Effectiveness ($/ton)
Overall Incremental
1
0
0
0
N/A
N/A
2
130
63
42
1,500
1,500
3
590
230
52
4,500
18,000
B-14
-------
Table B-9. National Fifth Year Impacts of Options for SO2 Limits Considered for
Modified and Reconstructed Small Sulfur Recovery Plants Subject to 40 CFR
Part 60, Subpart Ja
Option
Capital Cost
($1,000)
Total Annual Cost
($l,000/yr)
Emission Reduction
(tons S02/yr)
Cost-Effectiveness (S/ton)
Overall Incremental
1
0
0
0
N/A
N/A
2
1,600
670
380
1,800
1,800
3
7,800
2,600
470
5,700
23,000
The costs for Option 2 are reasonable considering the emission reductions achieved, but
the incremental impacts shown in Table 9 and Table 10 for Option 3 are beyond the costs that the
Agency believes are reasonable for these small units to achieve an additional 100 tons per year of
SO2 emission reductions. The additional equipment needed to achieve these reductions
quadruples the capital costs. These smaller units would also generally be found at small
refineries. Based on these projected impacts and available performance data, we support our
original determination that BDT is Option 2, or 99 percent sulfur recovery. For new SRP, this
option achieves SO2 emission reductions of 42 tons/yr from a baseline of 150 tons/yr at a cost of
$1,500 per ton of SO2. For modified and reconstructed SRP, this option achieves SO2 emission
reductions of 380 tons/yr from a baseline of 1,400 tons/yr at a cost of $1,800 per ton of SO2.
We note that we are also revising the format of the standard in response to public
comments in terms of sulfur outlet concentrations. Based on the Option 2 BDT selection of a
recovery efficiency of 99 percent, the emission limit for small SRP is either 2,500 ppmv SO2 or
3,000 ppmv reduced sulfur compounds and 100 ppmv of H2S, both of which are determined on a
dry basis, corrected to 0 percent O2.
B.7 NOx Limit for Process Heaters
Comment: Several commenters stated that the 80 ppmv NOx limit for process heaters is
not stringent enough. Commenters stated that considering recent settlement negotiations and
regulation development, NOx emissions reductions well below 80 ppmv can be achieved cost
effectively. The commenters stated that NOx emissions of less than 40 ppmv at 0 percent O2 are
achievable with combustion modifications such as LNB, ultra low-NOx burners (ULNB), and
flue gas recirculation technologies; post-combustion controls such as SCR, SNCR, and LoTOx™
achieve NOx reductions an order of magnitude below those from combustion modifications. The
B-15
-------
commenters noted that Bay Area Air Quality Management District (BAAQMD) Regulation 9,
Rule 10. requires process heaters to meet a 0.033 lb/MMBtu NOx limit (roughly 32 ppmv NOx
at 0 percent oxygen). One commenter stated that 30 ppmv has been demonstrated under consent
decrees to be an achievable level and ample technology exists. The commenters also noted that
7 to 10 ppmv NOx limits (at 3 percent oxygen) have been achieved in practice. One commenter
stated that NSPS subparts J and Ja should impose NOx emission limits on all fuel gas
combustion devices that are at least as stringent as the most stringent consent decree. Some
consent decrees require next generation ULNB designed to achieve NOx emissions rates of
0.012 to 0.020 lb/MMBtu (12 to 20 ppmv NOx at 0 percent oxygen). Commenters
recommending more stringent requirements suggested limits ranging from 7 ppmv NOx (at 3
percent oxygen) to 30 ppmv for new process heaters fueled by refinery fuel gas.
Other commenters stated that alternative monitoring options should be provided to small
fuel gas combustion devices due to the high costs of CEMS relative to the emissions from the
small devices. One commenter suggested an exemption from the fuel gas monitoring
requirements for process heaters less than 50 MMBtu/hr. Another commenter recommended an
exemption from the fuel gas monitoring requirements for process heaters less than 40 MMBtu/hr
as used by South Coast Air Quality Management District (SCAQMD).
Response: We revisited the BDT determination based on the public comments and
revised the methodology used to calculate the cost and emission reduction impacts for the
proposed standards. We evaluated three options as part of the BDT determination. Each option
consists of a potential NOx emission limit and applicability based on process heater size. These
differ slightly from the proposal options based on commenter suggestions. Option 1 would limit
NOx emissions to 80 ppmv or less for all process heaters with a capacity greater than 20
MMBtu/hr (the proposed standards). Option 2 would limit NOx emissions to 40 ppmv or less
for all process heaters with a capacity greater than 40 MMBtu/hr. This option is similar to many
consent decrees that set an emission limit of 0.040 lb/MMBtu (roughly 40 ppmv NOx at
0 percent oxygen) for process heaters greater than 40 MMBtu/hr. Option 3 would limit NOx
emissions to 20 ppmv or less for all process heaters with a capacity greater than 40 MMBtu/hr.
In each option, the NOx concentration is based on a 24-hour rolling average.
B-16
-------
The estimated fifth-year emission reductions and costs for each option for new process
heaters are summarized in Table B-10; impacts for modified and reconstructed process heaters
are summarized in Table B-l 1. Similar to the proposal analysis, we considered LNB, ULNB,
flue gas recirculation. SCR, SNCR, and LoTOx™ as feasible technologies. We believe that
nearly all process heaters at refineries that will become subject to subpart Ja can meet Option 1
or Option 2 using combustion controls (LNB or ULNB). Most process heaters would need to use
more efficient control technologies, such as LoTOx™ or SCR, to meet the NOx concentration
limit in Option 3. Per commenters' request to focus on the larger units, Options 2 and 3 do not
include process heaters between 20 MMBtu/hr and 40 MMBtu/hr. We evaluated the cost-
effectiveness of NOx control options for these units to achieve the proposed standard of 80
ppmv. For these process heaters with smaller capacities we found the cost-effectiveness ranged
from $3,500/ton to $4,200/ton of NOx reduced, which was determined not to be reasonable for
these small heaters, which would primarily be located at small refineries.
Table B-10. National Fifth Year Impacts of Options for NO, Limits Considered for New
Process Heaters Subject to 40 CFR Part 60, Subpart Ja
Option
Capital Cost
(SI,000)
Total Annual Cost
($l,000/yr)
Emission Reduction
(tons NOx/yr)
Cost-Effectiveness ($/ton)
Overall Incremental
1
9,000
7,300
4,800
1,500
1,500
2
9,000
7,500
5,200
1,400
500
3
110,000
30,000
5,900
5,100
37,000
Table B-l 1. National Fifth Year Impacts of Options for NO* Limits Considered for
Modified and Reconstructed Process Heaters Subject to 40 CFR Part 60,
Subpart Ja
Option
Capital Cost
(SI,000)
Total Annual Cost
($l,000/yr)
Emission Reduction
(tons NOj/yr)
Cost-Effectiveness ($/ton)
Overall Incremental
1
12,000
4,000
2,100
1,900
1,900
2
14,000
4,300
2,200
1,900
2,100
3
64,000
15,000
2,500
5,900
39,000
Based on the impacts in Tables B-10 and B-l 1, the costs of Options 1 and 2 are
reasonable compared to the emission reductions. The incremental cost between Options 2 and 3
B-17
-------
of almost $40.000/ton of NOx is not commensurate with the additional 1,000 tons of emission
reduction achieved for new and modified or reconstructed process heaters. Moreover, the capital
costs of Option 3 are about Si 50 million greater than the capital costs for Option 2, which are
only $23 million. Therefore, we conclude that BDT for process heaters greater than 40
MMBtu/hr is technology that achieves an outlet NO* concentration of 40 ppmv or less, or
Option 2. For new process heaters, this option achieves NOx emission reductions of 5,200 tons/yr
from a baseline of 7,500 tons/yr at a cost of $1,400 per ton of NOx. For modified and
reconstructed process heaters, this option achieves NOx emission reductions of 2,200 tons/yr
from a baseline of 3,200 tons/yr at a cost of $1,900 per ton ofNOx. Although we agree that lower
NOx concentrations are achievable, we determined that the incremental cost to achieve these
lower NOx concentrations was not reasonable.
B.8 Fuel Gas Combustion Devices
Comment: Several commenters contended that the proposed standards for fuel gas
combustion devices were not stringent enough; EPA should ensure that the best demonstrated
emission control technologies are installed as the industry is modernized. Given the significant
hazards to human health and the environment posed by SO2 emissions, the commenters
suggested that the 365-day average limits should be 40 ppmv TRS and 5 ppmv SO2. The
commenters also recommended that EPA tighten the 3-hour concentration limit to 100 ppmv
TRS. On the other hand, another commenter contended that although amine treatment
applications for product gases can achieve H2S concentrations of 1 to 5 ppmv, a tighter standard
is not BDT for refinery fuel gas.
Several commenters objected to the addition of the 60 ppmv H2S and 8 ppmv SO2 limits
(365-day rolling average) in the proposed subpart Ja standards for fuel gas combustion devices
because they are infeasible and/or not cost-effective. According to commenters, EPA
erroneously assumed that the additional reductions could be achieved with existing equipment.
Although this may be true in some cases, commenters asserted that some refineries would need
to add additional amine adsorber/regenerator capacity and some may also need to add additional
sulfur recovery capacity (e.g.. an additional Claus train and tail gas treatment unit). One
commenter requested an exemption be provided for refineries that cannot meet the tighter long-
term standard by simply increasing their amine circulation rates. One commenter stated that
there will .be little incremental environmental benefit from the long-term limit, and it
unnecessarily penalizes refineries that designed their amine systems to treat to levels near the
B-18
-------
proposed annual standard. The commenters provided cost data for examples of projects
requiring new amine adsorption units to show that the proposed standards are not cost-effective.
A number of commenters particularly opposed the proposed revision to include TRS
limits for fuel gas produced from coking units or any fuel gas mixed with fuel gas produced from
coking units. One commenter noted that some State and local agencies have specific TRS
standards, but these requirements were not based on a BDT assessment. According to
commenters. EPA has included no technical basis for the achievability of the TRS fuel gas
standard or explanation of why control of TRS is limited to fuel gas generated by coking units.
The commenters recommended that EPA postpone adoption of a TRS limit until it has gathered
and evaluated adequate data to conclude that the limit is technically feasible and cost effective.
Commenters stated that EPA did not address the cost-effectiveness and non-air quality
impacts of the TRS standards and did not define BDT for the removal of TRS. One commenter
stated that without an established de minimis level, an entire fuel gas system could be subject to
the TRS limits if any amount of coker gas enters the fuel gas system. Amine scrubbing systems
are selective to H2S and are not suitable to other TRS compounds such as mercaptans, according
to the commenters. Commenters stated that the non-F^S TRS compounds are not amenable to
amine treating and there is no technology readily in-place at refineries for reducing non-I-hS TRS
compounds. Therefore, according to the commenters, removing these other TRS compounds
would require significant capital outlay for new equipment, costs that were not considered in the
impacts analysis.
One commenter provided an example of a treatment system installed to meet a facility-
wide fuel gas total sulfur standard of 40 ppmv; the commenter estimated the capital cost of the
entire system to be $150 million. The commenter also indicated that low-BTU gas from
flexicoking units would need to be specially treated at a capital cost of $61 million to achieve a
total sulfur content of less than 150 ppmv, and the treatment would increase energy consumption,
resulting in increases in NOx and CO emissions. Another commenter provided an order-of-
magnitude engineering estimate of $50 million to treat TRS down to 45 ppmv (long-term
average). Based on one commenters experience with a new fuel gas treating facility, non-acidic
TRS cannot be treated down to the proposed levels utilizing Merox-amine treatment. A cost-
effective solution could be natural gas blending at the affected combustion device; however, this
B-19
-------
option has the negative effect of reducing the production of refinery fuel gas and therefore
reducing the refinery's capacity for making gasoline.
Several commenters stated that the original BDT determination was based on amjne
scrubbing of H2S and not on SO2; the SO2 standard was simply a compliance option that was
calculated to be equivalent to the H2S concentration limit at 0 percent excess air. They also
asserted that EPA cannot use the SO2 option as a basis for the TRS standard because the SO2
option is not BDT. On the other hand, one commenter requested that EPA clarify the fuel gas
standards in subpart J to expressly indicate that the 20 ppmv SO2 limit is a valid compliance
option (instead of including it only in the monitoring section). According to the commenter,
focus has been on H2S due to the structure of the requirements of subpart J and permits rarely
require that combustion sources demonstrate compliance with the 20 ppmv SO2 limit. The
commenter stated that refiners clearly should be allowed to comply with the broader, more
comprehensive SO2 limit.
A few commenters noted that, as H2S is part of TRS, the TRS standard is even more
stringent than the H2S standard. One commenter recommended that no change in the fuel gas
standards be made or that the standards focus on H2S only with an alternative emission limit for
SO2. One commenter stated that EPA developed the 160 ppmv H2S standard to be more
stringent than the 20 ppmv SO2 standard specifically because H2S did not represent all of the
sulfur in the fuel gas. Commenters stated that using an F-factor approach (Method 19, 40 CFR
part 60, Appendix A-7), the TRS limit that is equivalent to the 20 ppmv SO2 emission limit is
260 ppmv and the TRS limit that is equivalent to the 8 ppmv SO2 emission limit is 104 ppmv.
Response: We initially concluded that fuel gas generated by the coking unit was mixed
with other fuel gases that were mostly H2S and that increasing the amine circulation rate would
result in additional H2S removal that could be used to meet the proposed standard. However,
based on a review of the available data, non-I^S sulfur content in coker fuel gas may be 300 to
500 ppmv. At these levels, specific treatment to reduce these other sulfur compounds would be
needed. As indicated by one commenter, a plant-wide total sulfur limit of 40 ppmv has been
achieved in practice in at least one refinery using a treatment train consisting of a Merox system,
sponge oil absorbers, MEA absorbers, and caustic wash towers. Therefore, total sulfur fuel gas
treatment methods are demonstrated. We evaluated the cost of this treatment based on
information provided in the public comments.
B-20
-------
Based on the public comments and additional data, we revisited the BDT determination
and assessed three options for increasing SO2 control of fuel gas combustion devices: (1) 20
ppmv SO2 or 162 ppmv H2S averaged over 3 hours; (2) Option 1 plus 8 ppmv SO2 or 60 ppmv
H2S averaged over 365 days; and (3) a compliance option of 162 ppmv TRS averaged over 3
hours and 60 ppmv TRS averaged over 365 days for fuel gas combustion devices combusting
fuel gas generated by a coking unit and Option 2 for combustion devices combusting fuel gas not
generated by a coking unit. Option I includes the same limits that are in subpart J. so there are
no additional costs or emission reductions beyond those expected from the application of subpart
J. To address the commenters' concerns that not all facilities have available amine capacity to
ensure compliance with the new long-term limits, we revised our proposal analysis to include
additional costs for the estimated 10 percent of the affected facilities that would increase their
amine capacity to achieve Option 2. We estimated costs for a separate treatment train that can
treat TRS for Option 3 because, based on the public comments received, we have concluded that
amine treatment systems are not effective for non-h^S components of TRS.
The estimated fifth-year impacts of each of these options for new fuel gas combustion
devices are presented in Table B-12; the impacts for modified and reconstructed fuel gas
combustion devices are presented in Table B-13.
Table B-12. National Fifth Year Impacts of Options for SO2 Limits Considered for New
Fuel Gas Combustion Devices Subject to 40 CFR Part 60, Subpart Ja
Capital Cost
Total Annual Cost
Emission Reduction
Cost-Effectiveness ($/ton)
Option
($1,000)
($l,000/yr)
(tons SOj/yr)
Overall
Incremental
1
0
0
0
N/A
N/A
2
1,200
770
520
1,500
1,500
3
100,000
13,000
930
14,000
31,000
Table B-13. National Fifth Year Impacts of Options for SO2 Limits Considered for
Modified and Reconstructed Fuel Gas Combustion Devices Subject to 40 CFR
Part 60, Subpart Ja
Capital Cost
Total Annual Cost
Emission Reduction
Cost-Effectiveness ($/ton)
Option
($1,000)
($l,000/yr)
(tons S02/yr)
Overall
Incremental
1
0
0
0
N/A
N/A
2
33,000
11,000
4,700
2,400
2,400
3
1,700,000
200,000
9,100
22,000
42,000
B-21
-------
Overall costs for Options 1 and 2 are reasonable compared to the emission reduction
achieved for new. modified and reconstructed fuel gas combustion devices. We further
evaluated the incremental costs and reductions between the three options and found that they
were reasonable for Options 1 and 2, while the incremental cost for Option 3 is not. While
Option 3 provides significant additional SO2 emission reductions, the additional capital cost of
$1.7 billion is high and could pose a significant barrier to future refinery upgrades and
expansions. Based on these impacts and consideration of current operating practices, we
conclude that BDT is use of technology that reduces the emissions from affected fuel gas
combustion devices to 20 ppmv SO2 or 162 ppmv H2S averaged over 3 hours and 8 ppmv SO2 or
60 ppmv H2S averaged over 365 days, or Option 2. For new fuel gas combustion devices, this
option achieves SO2 emission reductions of 510 tons/yr from a baseline of 1,000 tons/yr at a cost
of $ 1,400 per ton of SO2. For modified and reconstructed fuel gas combustion devices, this
option achieves SO2 emission reductions of4,700 tons/yr from a baseline of 10,000 tons/yr at a
cost of $2,400 per ton of SO2.
We note that although we have determined that Option 3 is not BDT and we will not limit
the amount of SO2 emissions from combustion of sulfur compounds other than H2S in subpart Ja,
we plan to continue to work with the industry to understand the magnitude of these SO2
emissions and to identify technologies that can be cost effectively applied to reduce the
emissions. We have learned through this process that the SO2 emissions from combustion of
TRS in coker gas are generally not reflected in emission inventories and we plan to explore this
issue in greater detail in the future to determine where SO2 emissions are underestimated and the
best way to correct the inventories.
Comment: Several commenters stated that it is impossible for a refinery owner or
operator to specify, acquire, install, and calibrate a continuous monitoring system within 15 days
of a change that increases the H2S concentration such that an exempt stream is no longer exempt.
One commenter suggested quarterly stain tube sampling for 1 year prior to revoking an
exemption from monitoring to confirm the change is permanent. The commenter suggested that
after 1 year of confirmation, an additional 12 months be provided to specify, acquire, install, and
calibrate the continuous monitoring system. One commenter suggested 1 year be provided for
installing a CEMS, while another commenter suggested 180 days be provided (with an allowance
for an additional extension) for installing a CEMS, rather than the 15 days proposed. .
B-22
-------
Response: We believe that in most cases, the process change would be a deliberate,
planned act and that the potential consequences of this deliberate change would be evaluated.
That is, before the equipment is modified, the refinery owner or operator is expected to assess the
impacts of this change on the exempted fuel gas stream. If the change is expected to increase the
sulfur content of the fuel gas, than the owner or operator can plan to install the required CEMS
when modifying the process. We recognize that some process changes may have unexpected
consequences, and a modification that was not expected to increase the sulfur content of the fuel
gas can result in an increase in sulfur content. In this case, it may be impossible to install the
required CEMS within 15 days. However, quarterly sampling does not provide any basis by
which the refinery owner or operator can demonstrate compliance with the H2S concentration
standard. Instead, we have added provisions that require an owner or operator to install a CEMS
as soon as practicable and no later than 180 days after a change that makes the stream no longer
exempt. Between the process change and the time a CEMS is installed, the owner or operator
must conduct daily stain tube sampling to demonstrate compliance with the H2S concentration
standard. During this time, a single daily sample exceeding 162 ppmv must be reported as an
exceedance of the 3-hour H2S concentration limit and a rolling 365-day average concentration
must be determined. A daily average H2S concentration of 5 ppmv is to be used for the days
prior to the process change for the previously exempt stream in calculating the rolling 365-day
average concentration.
B.9 Flaring of Refinery Fuel Gas
Comment: Several commenters supported the proposed work practice standards to
eliminate routine flaring and develop startup, shutdown, and malfunction (SSM) plans; the
commenters opposed the co-proposal of no standards. One commenter supported the
determination that elimination of routine flaring is BDT, citing reductions in hydrocarbon, NO\,
SO2, and carbon dioxide (CO2) emissions. One commenter stated that both subparts J and Ja
should explicitly require that flaring be used only as a last resort in unusual circumstances, such
as emergencies, and not on a routine basis. Commenters asserted that monitoring on an ongoing
basis is needed to verify that no flaring of nonexempt gases occurs. Commenters stated that
subpart Ja should also require refiners to install a flare gas recovery system, although such
B-23
-------
requirements should not preclude monitoring requirements. One commenter stated that the
NSPS should require a SSM plan to eliminate venting or flaring during such planned start-up,
shutdown, and maintenance activities and explicitly prohibit venting or flaring during these
planned activities; proper operation and maintenance practices should completely eliminate the
need to use flares during these activities. One commenter noted that those refineries that have
evaluated their startup and shutdown procedures to reduce or eliminate direct venting or flaring
during planned startup and shutdown events have demonstrated the best technology; therefore,
their actions represent BDT and should be adopted in the NSPS. The commenters also supported
conducting a root cause analysis (RCA) in the event of flaring and other venting releases of 500
lb/day S02.
A number of commenters generally supported the intent to reduce flaring and the idea of
SSM plans to address flaring during planned startups and shutdowns (one commenter also
included combustion of high sulfur-containing fuel gases during a malfunction), flare
management plans, and RCA for flare events in excess of 500 lb/day. However, they opposed
the work practice standard for elimination of routine flaring and the proposed creation of fuel gas
producing units for subpart Ja. The commenters stated that the definition of "fuel gas producing
unit" is overly broad, making it difficult to determine what constitutes a modification or
reconstruction, and the proposed work practice standard for these units is infeasible,
unnecessary, and not cost-effective. Facility operators and regulators would have difficulty
discerning if a flaring event was caused by an affected fuel gas producing unit or a unit not
subject to the standard. One commenter indicated that there is no de minimis level by which
units that produce insignificant quantities of fuel gas can be excluded from the extensive work
practice standards.
Commenters recommended that the affected source be the flare which is already subject
to the standard as a fuel gas combustion device. The'commenters suggested that for each
affected flare, the facility would develop a written Flare Management Plan designed to minimize
flaring of fuel gas during all periods of operation. This plan, along with the RCA, would ensure
that all flaring events with potential excess emissions will be minimized. One commenter noted
that EPA could require a flare management plan for any flare tied to a fuel gas system that has an
affected fuel gas combustion device as a better alternative to "fuel gas producing units." One
commenter noted that an exemption from the notification requirements for modified or
B-24
-------
reconstructed units could be provided as an incentive for early adoption of the flare management
plan; another commenter suggested that regulatory incentives such as exemptions from
monitoring and developing flare management plans should be provided for facilities that have
installed flare gas recovery systems. One commenter supported this type of requirement for
flares currently subject to subpart J, assuming a minimum of 9 months is provided for plan
development and implementation. On the other hand, one commenter noted that the definitions
of the affected facility under subparts J and Ja are different and recommended that the distinction,
be made stronger so that it is clear that existing process unit facilities are "grandfathered" and
exempt from the flaring minimization standards.
One commenter suggested that the work practices language should be clarified to indicate
that routing offgas to the flare system would be acceptable if the system was equipped with a
flare gas recovery system. The prohibition should be specific to the flare itself as some flare
systems are equipped with recovery compressors, the use of which should be encouraged rather
than discouraged.
Comment'ers stressed the need for flares as safety devices; any flare minimization
program must not interfere with the ability of the refinery owner or operator to use flares for
safety reasons. The commenters stated that "routine" flaring cannot be adequately defined in
practice; therefore, restrictions on "routine" flaring will lead to unsafe operations in attempts to
avoid enforcement actions. The commenters requested that EPA include language in the
regulation, consistent with the preamble discussion, that: "Nothing in this rale should be
construed to compromise refinery operations and practices with regard to safety."
One commenter indicated that the proposed work practice standards for "no routine
flaring" interfere with flare minimization plans implemented in response to consent decrees. The
proposed work practice standard could be interpreted as prohibiting flaring during start-up and
shutdown, and EPA has not determined this to be BDT. The commenter stated that the
BAAQMD analysis applies to eliminating flaring during normal operation [similar to proposed
§60.103a(b)], not during start-up and shutdown as in proposed §60.103a(a). The commenter
provided cost estimates for one refinery to install a recovery system to eliminate flaring during
start-up and shutdowns; the costs ranged from $200,000 to $800,000 per ton of VOC reduced
and higher for other criteria pollutants. Therefore, they contend §60.103a(a) should clearly
exclude start-up and shutdown gases.
B-25
-------
A few commenters provided overall project costs for flare gas recovery projects
indicating the annual costs are higher than those in the analysis supporting the proposed work
practice. One commenter stated that EPA underestimated the cost of flare gas recovery systems
and, given the uncertainty in emission reductions, contended that flare gas recovery systems for
the no-flaring option are not cost-effective within the NSPS context. The commenter also stated
that the regulation should include maintenance provisions for flare gas recovery systems (that
allow flaring) during times of routine and non-routine maintenance, as no redundant capacity
within the flare system exists.
A number of commenters provided an alternative to EPA's proposed work practice
standards. The suggestions included a 500 lb/day SO2 standard tied with a flare management
plan as an alternative compliance option (to the H2S concentration limit) for flares. The
commenters recommended that this alternative compliance option be provided in both subparts J
J
and Ja and noted that it could be used as an incentive for the flare management plan to cover all
flares. One commenter also noted that these requirements should be applicable to flares that
receive process gas, fuel gas, or process upset gas; they should not be applicable to flares used
solely as an air pollution control device, such as a flare used exclusively to control emissions
from a gasoline loading rack. Another commenter clarified that if the refinery elects to comply
with this alternative for any flare, all flares at the refinery would need a flare management plan.
The commenter noted that EPA could choose to set the 500 lb/day SO2 limit as a total for all
flares for which the alternative compliance option is chosen (i.e.. if the alternative compliance
option is selected for two flares at a refinery, the total emissions from both flares would be
limited to 500 lb/day).
Response: Although commenters suggested that certain provisions be made applicable to
facilities subject to subpart J, the following provisions are only applicable to facilities subject to
subpart Ja as CAA section 111 provides that new requirements apply only to new sources. We
considered these comments and agree that the standards are more straightforward when the
affected facility is defined as the flare. Therefore, we have eliminated "fuel gas producing units"
as an affected facility in this final rule, and we specifically define a flare as a subset of fuel gas
combustion device, which is an affected facility in this final rule. A "flare" means "an open-
flame fuel gas combustion device used for burning off unwanted gas or flammable gas and
liquids. The flare includes the foundation, flare tip, structural support, burner, igniter, flare
B-26
-------
controls including air injection or steam injection systems, flame arrestors. knockout pots, piping
and header systems."
There are three general work practice standards that were proposed for "fuel gas
producing units," which may be summarized as follows: (1) the "no routine flaring"
requirement; (2) flare minimization plan for start-up, shutdown, and malfunction events; and (3)
a root-cause analysis for SO2 releases exceeding 500 lb/day (which was proposed for all affected
fuel gas producing units). The "no routine flaring" work practice was not intended to prohibit
flaring during SSM events; the provisions were intended to apply only, during normal operating
conditions. We agree with the commenter that suggested that nothing in this rule should be
construed to compromise refinery operations and practices with regard to safety. Additionally,
as discussed in the preamble to the proposed rule, we specifically rejected a prohibition on
flaring for planned start-up and shutdown events. We agree with the commenters that noted that
numerous refineries have demonstrated that flare minimization during planned start-up and
shutdown activities can greatly reduce flaring during these events. We do believe, however, that
a complete elimination of flaring during these events is very site-specific and although it is
reported to have been achieved at a limited number of refineries, we do not have information to
suggest that it has been adequately demonstrated for universal application. As "no routine
flaring" is difficult to define in practice, we have re-evaluated BDT using more specific options.
Option 1 is no additional standards for flares. In Option 2, any routine emissions event or
any process start-up, shutdown, upset or malfunction that causes a discharge into the atmosphere
more than 500 pounds per day of SO2 (in excess of the allowable emission limit) from an
affected fuel gas combustion device or sulfur recovery plant would require a root cause analysis
to be performed. This approach is similar to what is included in most consent decrees. We are
also including a requirement for continuous monitoring of TRS for all gases flared (including
those from upsets, startups, shutdowns, and malfunction events), in order to accurately measure
SO2 emissions from affected flares.
Option 3 includes: (1) the SO2 root cause analysis in Option 2; (2) a limit on the fuel gas
flow rate to the flare of 250,000 scfd; and (3) a flare management plan for SSM events. The
flow limit of 250,000 scfd is based on our cost analysis that indicates that for typical gas streams
in quantities above this limit, the value of recovered fuel completely offsets the costs of installing
and operating recovery systems. Many refineries have implemented flare gas recovery to reduce
B-27
-------
energy needs and save money. The flare management plan must: (I) include a diagram
illustrating all connections to each affected flare; (2) identify the flow rate monitoring device and
a detailed description of manufacturer's specifications regarding quality assurance procedures;
(3) include standard operating procedures for planned start-ups and shutdowns of refinery
process units that vent to the flare (such as staging of process shutdowns) to minimize flaring
during these events; (4) include procedures for a root cause analysis of any process upset or
equipment malfunction that causes a discharge to the flare in excess of 500,000 scfd; and (5)
include an evaluation of potential causes of fuel gas imbalances (i.e., excess fuel gas), upsets or
malfunctions and procedures to minimize their occurrence and records to be maintained to
document periods of excess fuel gas. Excess emission events for the flow rate limit of 250,000
scfd and the result of root cause analysis must be reported in the semi-annual compliance reports.
Option 4 is identical to Option 3 except that flaring is limited to 50,000 scfd. This level
is estimated to be a baseline level that accounts for the flow requirement needed to maintain safe
operations of the flare (i.e.. flow of sweep gas and compressor cycle gas). For both Option 3 and
Option 4, the limit on the flow rate does not apply during malfunctions and unplanned startups
and shutdowns. The flow rate limits in Options 3 and 4 were developed to reduce VOC, SO2,
and NOx emissions; the limits are based on 30-day rolling average flow rate values. '
It is anticipated that a flare gas recovery system will be used to comply with Options 3
and 4 when a flare is currently used on a continuous basis, and the recovered flare gas offsets
natural gas purchases. The cost-effectiveness of the flare gas recovery system is primarily
dependent on the quantity of gas that the system can recover. Many refineries have already
implemented similar work practices through consent decrees and local rules (BAAQMD and
SCAQMD), and these requirements have had a demonstrated reduction in flaring events. Flare
gas recovery will reduce SO2, NOx, and VOC emissions. However, if a refinery produces more
fuel gas than the refinery needs to power its equipment, there is no place the refinery can use the
recovered fuel gas and there is no additional natural gas purchases to offset. In these cases, flare
gas recovery is not considered technically feasible because the excess fuel gas will have to be
flared. Therefore, we have included specific provision within the flare management plan to
address instances of excess fuel gas. For periods when the refinery owner or operator can
demonstrate, through records of natural gas purchases or other means as described in their flare
B-28
-------
management plan, that the refinery is fuel gas rich, compliance with the flow limit is
demonstrated by implementing the procedures described in the flare management plan.
Impacts for each of the four options are based on estimates of current flaring quantities
and include the root cause analysis, flare management plan, and flare gas recovery systems when
needed. The impacts for each option for new flares are presented in Table B-14; impacts for
modified and reconstructed flares are presented in Table B-l 5.
Table B-14. National Fifth Year Impacts of Options for Work Practices Considered for
New Flaring Devices Subject to 40 CFR Part 60, Subpart Ja
Capital
Total Annual
Emission
Emission
Emission
Cost-Effectiveness (S/ton)
Cost
Cost
Reduction
Reduction
Reduction
Option
($1,000)
($l,000/yr)
(tons S02/yr)
(tons NO,/yr)
(tons VOC/yr)
Overall
Incremental
1
0
0
0
0
0
N/A
N/A
2
0
23
15
0
0
1,600
1,600
3
8,800
(1,300)
16
1
41
(23,000)
(31,000)
4
15.000
(840)
16
1
52
(12,000)
43,000
Table B-l 5. National Fifth Year Impacts of Options for Work Practices Considered for
Modified and Reconstructed Flaring Devices Subject to 40 CFR Part 60,
Subpart Ja
Option
Capital
Cost
($1,000)
Total
Annual Cost
($l,000/yr)
Emission
Reduction
(tons S02/yr)
Emission Emission
Reduction Reduction
(tons NO,/yr) (tons VOC/yr)
Cost-Effectiveness (S/ton)
.Overall Incremental
1
0
0
0
0 0
N/A
N/A
2
0
92
59
0 0
1,600
1,600
3
35,000
(5,300)
64
4 165
(23,000)
(31,000)
4
59,000
(3,300)
66
6 207
(12,000)
43,000
Based on these impacts and consideration of technically feasible operating practices, we
conclude that BDT is Option 3. Option 3 includes a set of work practice standards that requires
root cause analysis for a discharge into the atmosphere in excess of 500 pounds per day of SO2
(over the allowable emission limit) from a fuel gas combustion device or sulfur recovery plant or
in excess of 500,000 scfd flow to a flare. It also includes a flare management plan. Finally, fuel
gas flow to the flare is limited to 250,000 scfd. To support implementation of these
requirements, monitoring and reporting of the flow rate and sulfur content is required. For new
flaring devices, this option achieves SO2 emission reductions of 16 tons/yr from a baseline of 32
B-29
-------
tons/yr, NOx emission reductions of 1 tons/yr from a baseline of 2 tons/yr, and VOC emission
reductions of 41 tons/yr from a baseline of 67 tons/yr with a net fuel savings of $23,000 per ton
of combined SO2, NOx, and VOC. For modified and reconstructed flaring devices, this option
achieves SO2 emission reductions of 64 tons/yr from a baseline of 129 tons/yr, NOx emission
reductions of 4 tons/yr from a baseline of 7 tons/yr, and VOC emission reductions of 165 tons/yr
from a baseline of 266 tons/yr with a net fuel savings of $23,000 per ton of combined SO2, NOx,
and VOC.
Comment: Several commenters requested clarification of how one would assess a flare
"modification." Questions included: (1) how the emission basis of a flare should be calculated;
(2) if the modification determination would be based on flare capacity or increase in discharge
capability of units connected to the flare; (3) whether the modification determination would
include all possible flaring events or just non-emergency flaring; (4) whether adding a new line
to a flare is considered to increase the capacity of the flare and cause a modification; (5) whether
flare tip replacements are considered routine maintenance instead of a modification of the flare,
even if the new flare tip has a different geometry (e.g.. a larger diameter to reduce noise); and (6)
how SSM streams are considered when calculating baseline emissions for a modification
determination. The commenters also suggested that EPA should clarify whether and how the
exemption in §60.14(e)(2) applies to a flare, including how the production rate for a flare would
be defined.
Response: Section 60.14(a) defines modification as follows: "Except as provided in
paragraphs (e) and (f) of this section, 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 shall be considered a modification." Section 60.14(e) provides exclusions for
maintenance activities, increased production rates, increased hours of operation, etc. However,
except for the maintenance exclusion, the other exemptions are either not applicable or
ambiguous when applied to a flare. More importantly, §60.14(f) states that "Applicable
provisions set forth under an applicable subpart of this part shall supersede any conflicting
provisions of this section." Therefore, to eliminate ambiguity, we specifically define what
constitutes a flare modification in subpart Ja.
A flare is considered to be modified in one of two ways. First, a flare is considered to be
modified when any piping from a refinery process unit or fuel gas system is newly connected to
B-30
-------
the flare. This new piping could allow additional gas to be sent to the flare, consequently
increasing emissions from the flare. Second, a flare is considered to be modified if that flare is
physically altered to increase flow capacity.
B.10 Delayed Coking Units
Comment: Several commenters supported the proposal that requires delayed coking units
to depressure the coke drums to the fuel gas system down to 5 psig. One commenter supported
venting the delayed coker gas to a flare or to the atmosphere at pressures less than 5 psig; at
pressures greater than 5 psig, the commenter suggested that the rule should only prohibit gases
from being sent to a flare and allow any other disposition. That is, the commenter stated that
EPA should not restrict the disposition of the coker depressurization gas to only the fuel gas
system.
One commenter supported inclusion of a coke drum pressure limit above which the coke
drum exhaust gases must be sent to a recovery system, disagreed that it is technically infeasible
to divert emissions for recovery at pressures below 5 psig, and urged EPA to require venting
until the pressure drops below 2 psig. The commenter recently issued a permit including the 2
psig level, and although the modification has not been completed, the commenter believes the
requirement is technically feasible.
A number of commenters objected to the finding that BDT is to depressure delayed
coking units to the fuel gas system down to 5 psig. Commenters provided examples of coking
units whose current mode of operations (e.g.. set points or timed cycles) may divert to a flare or
to the atmosphere at pressures of approximately 10 to 20 psig and that it would not be cost-
effective to modify these units to comply with the proposed work practice standard. One
commenter supported the premise that it is cost-effective for delayed coking discharge to be
routed to fuel gas blowdown, but depressurization down to 5 psig may not be feasible with
existing equipment; the commenter recommended that the work practice simply require a closed
blow down system following procedures described in the facility's SSM plan. At a minimum, an
alternative is needed for existing units that would require capital expenditure to meet the 5 psig
proposal. One commenter stated that compressors cannot recover blowdown system gases at
pressures below the fuel gas recovery compressor suction pressure. The minimum pressure at
which a suction compressor can operate depends on the design of the coking unit and the
blowdown management system. Because there is uncertainty surrounding the available emission
B-31
-------
information, the costs are not minimal in most cases, and the emissions are difficult to measure,
the commenter stated that EPA cannot determine that controls on coker vents is BDT.
Response: Based on the public comments, we re-evaluated BDT for delayed coking
units. We considered three options: (1) depressurization down to 15 psig; (2) depressurization
down to 5 psig; and (3) depressurization down to 2 psig. We estimated that the baseline is, on
average, depressurization down to 15 psig and then venting to the atmosphere. Therefore, there
are no impacts for Option 1. Impacts for Options 2 and 3 were estimated based on the baseline
conditions, the size of typical coke drums, and cost information provided in public comments.
We also collected emissions test data to support and verify the projected emissions and emission
reductions.
The impacts for each option for new delayed coking units are presented in Table B-16;
impacts for modified and reconstructed delayed coking units are presented in Table B-17.
Table B-16. National Fifth Year Impacts of Options for Work Practices Considered for
New Delayed Coking Units Subject to 40 CFR Part 60, Subpart Ja
Option
Capital Cost
($1,000)
Total Annual
Cost
($l,000/yr)
Emission
Reduction
(tons S02/yr)
Emission
Reduction
(tons VOC/yr)
Cost-Effectiveness ($/ton)
Overall Incremental
1
0
0
0
0
N/A
N/A
2
2,400
230
170
10 .
1,200
1,200
3
24,000
2,300
230
13
9,500
37,000
Table B-17. National Fifth Year Impacts of Options for Work Practices Considered for
Modified and Reconstructed Delayed Coking Units Subject to 40 CFR Part 60,
Subpart Ja
Option
Capital Cost
($1,000)
Total Annual
Cost
($l,000/yr)
Emission
Reduction
(tons S02/yr)
Emission
Reduction
(tons VOC/yr)
Cost-EfTectiveness ($/ton)
Overall Incremental
1
0
0
0
0
N/A
N/A
2
14,000
1,400
260
15
4,900
4,900
3
54,000
5,100
340
19
14,000
45,000
Based on these impacts and consideration of technically feasible operating practices, we
confirmed our conclusion at proposal that BDT is depressurization down to 5 psig, or Option 2.
For new delayed coking units, this option achieves SO2 emission reductions of 170 tons/yr from
B-32
-------
a baseline of 520 tons/yr and VOC emission reductions of 2 tons/yr from a baseline of 7 tons/yr
at a cost of $ 1,300 per ton of combined SO2 and VOC. For modified and reconstructed delayed
coking units, this option achieves SO2 emission reductions of 260 tons/yr from a baseline of 780
tons/yr and VOC emission reductions of 4 tons/yr from a baseline of 11 tons/yr at a cost of
$5,100 per ton of combined SO2 and VOC. Although Option 3 has been established in one
refiner's permit, this level of depressurization has not been demonstrated in practice.
Additionally, the difference in the quantity of gas released when the set point is 2 psig rather than
5 psig is relatively small, 80 tons of SO2 and 4 tons of VOC, and the resulting incremental cost-
\
effectiveness from Option 2 to Option 3 is about $40,000/ton, which is much greater. Therefore,
Option 3, or depressurization down to 2 psig, is not BDT.
B.ll Summary of Results for New Sources and Modified and Reconstructed Sources
Below in Table B-18 is a summary of results for the analyses done above for options
applied to new sources. Table B-19 contains a similar summary for the modified and
reconstructed sources.
B-33
-------
Table. B-18. National Incremental Emission Reductions and Cost Impacts for Options Applied to New Petroleum Refinery
Units Subject to Final Standards Under 40 CFR Part 60, Subpart Ja (Fifth Year After Proposal)
Process
- Unit
Pollutant
Controlled
Option
Total
Capital
Cost
($1,000)
Total
Annual Cost
($l,000/yr)
Annual
Emission
Reductions
(tons PM/yr)
Annual
Emission
Reductions
(tons SOz/yr)
Annual
Emission
Reductions
(tons NOx/yr)
Annual
Emission
Reductions
(tons
VOC/yr)
Annual
Cost-
Effectiveness
(S/ton)
Incremental
Annual Cost-
Effectiveness
($/ton)
FCCU
PM
1 (baseline)
2"
3,600
1,100
235
• 5,600
5,600
3 '
7,100
1,700
300
6,700
10,900
S02
1 (baseline)
2"
0
1,400
1,993
700
700
NOx
1
900
300.
368
900
900
2"
1,200
600
859
700
600
3
12,200
3,600
1,382
2,600
5,800
Small SRP
S02
1 (baseline)
2"
100
100
42
1,500
1,500
-
3
600
200
52
4,500
17,700
Fuel gas
combustion
devices
so2
1 (baseline)
2°
1,200
800
524
1,500
1,500
3
100,200
13,200
926
14,300
31,000
Process
heaters
NOx
1
9,000
7,300
4,841
1,500
1,500
-
2"
9,000
7,500
5,237
1,400
1,400
3
110,700
30,100
-
5,853
5,100
36,700
(continued)
-------
Table B-18. National Incremental Emission Reductions and Cost Impacts for Options Applied to New Petroleum Refinery
Units Subject to Final Standards Under 40 CFR Part 60, Subpart Ja (Fifth Year After Proposal) (continued)
Process Unit
Pollutant
Controlled
Option
Total
Capital
Cost
($1,000)
Total
Annual Cost
($l,000/yr)
Annual
Emission
Reductions
(tons PlW/yr)
Annual
Emission
Reductions
(tons S02/yr)
Annual
Emission
Reductions
(tons NOx/yr)
Annual
Emission
Reductions
(tons
VOC/yr)
Annual
Cost-
Effectiveness
($/ton)
Incremental
Annual Cost-
Effectiveness
($/ton)
Flare gas
minimization
so2. voc
1 (baseline)
2
0
23
15
0
0
1,600
1,600
3"
8,800
-1,300
16
I
41
-23,000
-31,000
4
15,000
-840
16
1
52
-12,000
43,000
Delayed
coking units
so2. voc
1
2"
2,400
200
174
10
1.200
1,200
3
24.000
2.300
227
13
9.500
36,900
Sulfur pits
SO;
1
2"
700
100
30
2900
2,900
3
1.300
200
31
5,600
114,000
" Denotes selected option. All costs are in 2006 dollars. 83.3% of the PM emissions are PM25.
-------
Table B-19. National Incremental Emission Reductions and Cost Impacts for Options Applied to Modified and Reconstructed
Petroleum Refinery Units Subject to Final Standards Under 40 CFR part 60, subpart Ja (Fifth Year After
Proposal)
Annual
Incremental
Annual
Annual
. Annual
Emission
Annual
Annual
Total
Total
Emission
Emission
Emission
Reductions
Cost-
Cost-
Process
Pollutant
Capital Cost Annual Cost
Reductions
Reductions
Reductions
(tons
Effectiveness
Effectiveness
Unit
Controlled
Option
($1,000)
($l,000/yr)
(tons PM/yr)
(tons S02/yr)
(tons NO,/yr)
VOC/yr)
(S/ton)
($/ton)
FCCU
PM
1 (baseline)"
2
75,200
11,900
690
20,700
20,700
• 3
101,100
15,500
808
23,000
36,500
S02
1 (baseline)
2°
0
1,600
2.400
700
700
NOx
1
2,800
1,000
856
1,200
900
2"
3,700
1,600
1,784
900
700
3 ,
44,800
11,500
3.234
3,600
6.800
FCU
PM, S02
1 (baseline)
2"
10,400
3,200
' 1,000
5,900
500
500
NOx
1 (baseline)
2
800
200
410
400
400
3"
3700
900
657
1,300
2,700
4
6,000
1,300
745
1.700
5,000
Small SRP
S02
1 (baseline)
>
2"
1,600
700
381
1,800
1,800
3
7,800
2.600
466
5,700
23.000
(continued)
-------
Table B-19. National Incremental Emission Reductions and Cost Impacts for Options Applied to Modified and Reconstructed
Petroleum Refinery Units Subject to Final Standards Under 40 CFR part 60, subpart Ja (Fifth Year After
Proposal) (continued)
Process Unit
Pollutant
Controlled
Option
Total
Capital Cost
(SI,000)
Total
Annual Cost
($l,000/yr)
Annual
Emission
Reductions
(tons PM/yr)
Annual
Emission
Reductions
(tons S02/yr)'
Annual
Emission
Reductions
(tons NO,/yr)
Annual
Emission
Reductions
(tons
VOC/yr)
Annual
Cost-
Effectiveness
($/ton)
Incremental
Annual Cost-
Effectiveness
(S/ton)
Fuel gas
combustion
devices
SO,
1 (baseline)
2"
32,900
11,300
4,700
2,400
2,400
3
1,674,000
198,100
9,100
21,700
42,300
Process
heaters
NOx
1
11,700
4,000
2,075
1,900
1,900
2°
14,000
4,300
2,244
1,900
2,100
3
64,100
14,800
2,509
5,900
39,400
Flare gas
minimization
S02, VOC
1 (baseline)
/
59
2
0
100
0
0
1,600
1,600
3°
35,000
-5,300
64
4
165
-23,000
-31,000
4
y
59,000
-3,300
66
6
207
-12,000
43,000
Delayed
coking units
SO,. VOC
1
2"
14,400
1,400
261
15
4,900
4,900
3
54,000
5,100
340
19
14,200
45,100
Sulfur pits
S02
1
2"
7,700
900
269
3,500
3.500
3
15,300
1,900
275
6,800
138,800
" Denotes selected option. All costs are in 2006 dollars. 83.3% of the PM emissions are PM2 5.
-------
SECTION 5
ECONOMIC IMPACT ANALYSIS: METHODS AND RESULTS
The EIA is designed to inform decision makers about the potential economic
consequences of a regulatory action. The analysis consists of estimating the social costs of a
regulatory program and the distribution of these costs across stakeholders (consumers and
producers). As defined in EPA's (2000) Guidelines for Preparing Economic Analyses,1 social
costs are the value of the goods and services lost by society resulting from using resources to
comply with and implement a regulation and reductions in output.
5.1 Market Model
EPA constructed partial equilibrium models of the national markets for five major
petroleum products (motor gasoline, jet fuel, distillate fuel oil, residual fuel oil, and liquefied
petroleum gases).2 These models were used to measure the economic consequences of the
regulatory program in the intermediate run (when some factors of production are fixed and others
are variable).3 Partial equilibrium models track the effects of regulatory action in a single market,
while ignoring interactions with other markets.
Each of the 5 intermediate-run market models uses a common analytic expression to
estimate how an increase in the per-unit (per-gallon) costs of producing a product will impact
that product's price (Berck and Hoffmann, 2002; Fullerton and Metcalfe, 2002). This expression
is presented in Equation 5.1. A full description for how it is derived and used is provided in
Appendix C.
' Supply Elasticity „ _ ,, _
Aprice ^ Supply Elasticity Demand Elasticity ^ er~ a on ost
This approach follows EPA guidelines for analyzing the economic impacts of a
regulatory program (EPA, 1999; EPA, 2000).
5.2 Model Baseline
Standard EIA practice compares and contrasts the state of a market with and without a
regulatory policy. EPA selected 2012, the fifth year after proposal, as the baseline year for the
analysis. Forecasts for the price and consumption of each petroleum product in 2012 were
1 These guidelines are under review by the Agency.
2 National market models were selected in order to be consistent with the national-level cost estimates provided by
the engineering cost analysis.
3 For a complete discussion of how the intermediate run is defined, please see the OAQPS Economic Analysis
Resource Document (EPA, 1999).
-------
collected from the Energy Information Administration's Annual Energy Outlook and reported in
Chapter 3 (Tables 3-19 and 3-20). However, these data had to be standardized for use in EPA;s
models.
First, the 2007 Annual Energy Outlook reports the price of petroleum products in terms
of 2005 dollars. However, compliance costs were estimated in terms of 2006 dollars. Therefore,
to ensure that common units were being used, petroleum product prices were converted to 2006
dollars by dividing the 2012 forecasted price in 2012 by the ratio of the Consumer Price Indices
(CPI) in 2006 and 2005.
Second, the 2007 Annual Energy Outlook reports the quantity of petroleum products
consumed in terms of barrels, while the price of petroleum products is reported in terms of
dollars per gallon. Therefore, to ensure that common units were being used, the number of
barrels produced each year was divided by 42 (the number of gallons in a barrel). A summary of
the baseline data used in each of the five market models after these adjustments were made is
reported in Table 5-1.
Table 5-1. Baseline Market Data: 2012
Liquefied
Motor
Distillate
Residual
Petroleum
Market
Gasoline
Jet Fuel
Fuel Oil
Fuel Oil
Cases
Price ($2006/per gallon)
$2.11
$1.40
$2.04
$1.06
$1.55
Quantity (billion gallons/per year)
149.67
31.04
71.83
12.36
34.10
Sources: 2012 Petroleum product price and consumption forecasts: U.S. Department of Energy, Energy Information
Administration (EIA). 2007. "Annual Energy Outlook." Available at . As obtained on January 21, 2007.
2005 and 2006 Consumer Price Indices: U.S. Department of Energy, Energy Information Administration
(EIA). 2008. "Short-Term Energy Outlook: Real Petroleum Prices." Available at . As obtained on April 21, 2008.
5.3 Model Parameters
\
An essential component of partial equilibrium models are supply and demand price
elasticities. These elasticities measure the responsiveness of producers and consumers to prices
changes and determine how the social costs of a regulatory program are distributed between the
two groups of stakeholders. Economic theory suggests consumers will bear a higher share of the
economic welfare losses if the supply of a petroleum product is more responsive to price changes
than is the demand for that product. A summary of the estimates of demand and supply
elasticities used in this analysis is provided in Table 5-2.
5-2
-------
Table 5-2. Estimates of Price Elasticity of Demand and Supply
Liquefied
Motor
Distillate
Residual Fuel
Petroleum
Market
Gasoline
Jet Fuel
Fuel Oil
Oil
Cases
Demand elasticity
-0.69
-0.15
-0.75
-0.68
-0.8
Supply elasticity
1.24
1.24
1.24
1.24
1.24
Sources: U.S. Environmental Protection Agency. 1995. Economic Impact Analysis for Petroleum Refineries
NESHAP. EPA-452/R-95-003, Final Report. Washington DC: Government Printing Office.
5.4 Results
Chapter 4 reports that the estimated change in total annualized costs resulting from the
regulatory program is a cost of approximately $31 million (measured in 2006 dollars). According
to the 2007 Annual Energy Outlook, the forecasted consumption of all petroleum products in
2012 is 8.08 billion barrels or 339.25 billion gallons. Assuming that the production process of all
petroleum products are equally affected this regulatory program is expected to result in a
$0.000091 per-gallon cost to consumers ($31 million / 339.25 billion gallons).4
Based on EPA's partial equilibrium analysis, the costs induced by this regulatory
program do not have a significant impact on market-level prices or quantities. The results of this
analysis are summarized in Table 5-3. As this table shows, prices for all products rise by less
than 1 penny (0.003%-0.006%). In addition, the quantity of each petroleum product produced
declines. Motor gasoline and distillate fuel face the largest absolute quantity reductions (2.8 and
1.4 million gallons, respectively, or 0.002%), while,residual fuel oil sees the largest proportional
decline in production (0.004%).
As a result of higher prices, consumers of petroleum products see a decline in surplus.
For example, consumers of motor gasoline lose $8.78 million of surplus. In addition, producers
also receive a smaller surplus as a result of higher production costs. In the case of motor
gasoline, producers lose $4.89 million. Total surplus losses for consumers and producers of
motor gasoline are estimated to be $13.67 million. Thus, the total annualized loss in surplus in
the markets analyzed, which is an estimate of the social cost of this NSPS, is $27.30 million or
slightly less than the total annualized compliance cost in these markets. In addition to the loss in
surplus for these petroleum products, an additional $3.7 million in costs will affect markets for
other petroleum products.
4 In addition to motor gasoline, jet fuel, distillate fuel oil, residual fuel oil, and liquefied petroleum gases, other
petroleum products will experience cost savings as a result of the regulatory program. These products include
asphalt, lubricants, road oil, petroleum coke and others. However, these impacts were not explicitly modeled for
this analysis.
5-3
-------
Table 5-3. Summary of Intermediate Run Economic Impacts by Petroleum Product: 2012
Motor Gasoline
Jet Fuel
Distillate Fuel Oil
Residual Fuel Oil
Liquefied Petroleum
Gases
Change in price
0.003%
0.006%
0.003%-
0.006%
0.004%
Less than a penny per
gallon
Less than a penny per
gallon
Less than a penny per
gallon
Less than a penny per
gallon
Less than a penny per
gallon
Change in quantity
-0.002%
-0.001%
-0.002%
-0.004%
-0.003%
(-2.8 million gallons per
year)
(-0.2 million gallons per
year)
(-1.4 million gallons per
year)
(-0.4 million gallons per
year)
(-0.9 million gallons per
year)
Welfare Impacts
(Smillion)
Welfare Impacts
(Smillion)
Welfare Impacts
($mil!ion)
Welfare Impacts
($million)
Welfare Impacts
($million)
Change in consumer surplus
-$8.78
-$2.53
-$4.10
-$0.73
-$1.89
Change in producer surplus
-$4.89
-$0.31
-$2.46
-$0.40
-$1.22
Change in total surplus
-$13.67
-$2.83
-$6.56
-$1.13
-$3.11
-------
SECTION 6
SMALL BUSINESS ANALYSIS
The Regulatory Flexibility Act (RFA) generally requires an agency to prepare a
regulatory flexibility analysis of any rule subject to notice and comment rulemaking
requirements under the Administrative Procedure Act or any other statute unless the agency
certifies that the rule will not have a significant economic impact on a substantial number of
small entities.1 This section begins by describing the data and methods used for performing this
small business analysis and end by reporting the results of the analysis.
6.1 Data and Methods for Flexibility Analysis
The impact of the rule on small businesses is assessed using the ratio of compliance costs
to the annual revenue of the ultimate parent company. This is known as the cost-to-sales ratio or
CSR and it can be computed using the following equation:
JjACC
CSR = — (6:1)
TRj
where
TACC = total annual compliance costs,
i = indexes the number of affected plants owned by company j,
n = number of affected plants, and
TRj = total annual revenue of a representative ultimate parent company j in each
industry
If the CSR is less than 1 %, then the regulatory program is considered to not have a
significant impact on the parent company in question. This approach assumes affected firms
absorb the control costs, rather than pass them onto consumers in the form of higher prices.
In Chapter 3, 25 small companies owning petroleum refineries classified as small
according. As previously discussed, small businesses in the petroleum refining industry (NAICS
code 324100) are defined for the purposes of this rule as having 1,500 or fewer employees.2
1 Small entities include small businesses, small organizations, and small governmental jurisdictions.
" Refer to http://www.sba gov;idc/groups/i)Lihlic/documenis/sba homenaae/serv sstd tablendf.pdf for more
information on SBA small business size standards.
6-1
-------
Table 6-1 duplicates sales employment data for these 25 small companies originally
reported in Chapter 3.
Table 6-1. Characteristics of Small Businesses in the Petroleum Refining Industry
Parent
Company Sales
Parent Company
Facility Name
Refineries (#) (SMillions)
Employment (#)
AGE Refining & Manufacturing
1 287
52
American Refining Group
1 350
310
Arabian American Development Co
1 80
118
Calcasieu Refining Co.
1 638
51
Calumet Specialty Products
3 1,641
350
Countrymark Cooperative, Inc.
1 87
300
Cross Oil & Refining Co. Inc.
1 49
110
CVR Energy Inc.
1 3,038
577
Foreland Refining Co.
1 56
100
Frontier Oil Corp
2 4,000
727
Gary-Williams Co
1 97
200
Goodway Refining LLC
1 3
18
Greka Integrated Inc
1 22
145
Gulf Atlantic Operations LLC
1 9
32
Holly Corp.
2 4,023
859
Hunt Refining Co.
3 4,871
1,100
Lion Oil Co.
1 247
425
Pelican Refining Co. LLC
1 29.
62
Placid Refining Inc.
1 1,400
200
San Joaquin Refining Co., Inc.
1 288
20
Somerset Oil Inc
1 55
150
Trigeant Ltd.
1 5
50
Western Refining, Inc.
4 4,200
416
World Oil Corp
1 277.3
475
Wyoming Refining Co.
1 340
107
We note here, that we inadvertently used a different small business size standard for small refiners in the proposed
NSPS. The small business analysis for the final rulemaking incorporates the correct SBA small business size
standard of 1,500 employees per ultimate parent refiner. The effect of this correction on the affected small refiner
universe is an increase of one small firm. There is no effect on our determination of no significant economic
impact on a substantial number of small entities (to be shown later) as a result of this correction.
6.2 Results of Small Business Analysis
As described in Chapter 4, the EPA estimates that small businesses will invest in two new
or modified process units during the five-year period of analysis. Investing in these process units
would require the small businesses to incur an average $1.5 million per facility in annualized
compliance cost and earn an average $0.6 million per facility in cost savings as a result of the
final NSPS—a net total annualized compliance cost of $910 thousand per facility.
6-2
-------
Literature on the petroleum refining industry was examined to identify and characterize
small firms likely to be affected by the rule, The Oil & Gas Journal's 2008 Worldwide
Construction Update survey catalogued over 40 refining construction projects that have been
announced in the United States. Among the companies announcing construction projects3, three
were identified by EPA as small businesses—Frontier Oil Corp, Holly Corp, and Placid Refining
Inc. EPA therefore estimates that.three small businesses will invest in new or modified process
units during the five-year period of analysis.
As indicated in Table 6-1, each of these companies earned over $1 billion in revenue in
the base year for this analysis (2006). Assuming that these three small businesses (out of 25 total
small businesses identified by EPA) are representative of the small businesses that will invest in
new or modified process units over the five year period of analysis, their cost to sales ratios
would be less than 1%. As a result, the final NSPS is not expected to have a significant impact
on small companies.
After considering the economic impact of today's action on small entities, I certify that
this action will not have a significant economic impact on a substantial number of small entities.
Of the affected entities, none are estimated to incur annualized compliance cost over 1% of sales.
Although this action would not have a significant economic impact on any small entities,
EPA nonetheless has tried to reduce the impact of this action on small entities by incorporating
specific standards for small sulfur recovery plants and streamlining procedures for exempting
inherently low-sulfur fuel gases from continuous monitoring. In addition, EPA has updated this
small business analysis to incorporate capacity data for small refiners provided in a comment by
the Ad Hoc Coalition of Small Business Refiners.
3 Construction projects included in the Oil & Gas Journal 's analysis include new, expanded, and upgraded
processes.
6-3
-------
SECTION 7
HUMAN HEALTH BENEFITS OF EMISSIONS REDUCTIONS
7.1 Calculation of Human Health Benefits
In order to estimate the human health benefits of reducing emissions from refineries
through this final rulemaking, EPA used the benefits transfer approach and methodology
described in EPA's benefits analysis the Technical Support Document (TSD)1 accompanying the
recent National Ambient Air Quality Standards (NAAQS) for Ozone.2 In that RIA, EPA applied
a benefits transfer approach to estimate the PM2 5 co-benefits resulting from reductions in
emissions of NOn; EPA is adapting that method to estimate the health benefits for the projected
emission reductions of PM2 5 precursor pollutants associated with this final rulemaking.
EPA did not perform an air quality modeling assessment of the emission reductions
resulting from installing controls on these refineries because of the time and resource constraints
and the limited value of such an analysis for the purposes of developing the regulatory approach
for this final rule. This lack of air quality modeling limited EPA's ability to perform a
comprehensive benefits analysis for this final rulemaking since our benefits model requires either
air quality modeling or monitoring data.
To estimate the human health benefits of emission reductions from refineries for this analysis
in the absence of modeling data, we used the studies from the PM NAAQS Regulatory Impact
Analysis (RIA)3 to generate benefit-per-ton values. These PM25 precursor pollutant benefit per-
ton estimates provide the total monetized human health benefits (the sum of premature mortality
and premature morbidity) of reducing one ton of PM2 5 and PM2 5 precursor emissions from a
specified source. These benefits estimates have been updated in the final NSPS to utilize the
mortality valuation estimates obtained in the expert elicitation study, as mentioned in the
proposal. In addition, we also include VOC benefit-per-ton estimates.4 EPA has used a similar
technique in previous RIAs, beginning with the PM NAAQS RIA (U.S. EPA, 2006). EPA has
requested the SAB to review the presentation of benefits estimates based on the mortality
1 U.S. EPA, 2008a. Technical Support Document: Calculating Benefit Per-Ton estimates, Ozone NAAQS Docket
#EPA-HQ-OAR-2007-0225-0284.
" U.S. EPA, 2008b. . Regulatory.Impact Analysis, 2008 National Ambient Air Quality Standards for Ground-level
Ozone, Chapter 6. Available on the Internet at http://www.epa.gov/ttn/ecas/regdata/RIAs/6-ozoneriachapter6.pdf.
3 U.S. EPA, 2006. Regulatory Impact Analysis, 2006 National Ambient Air Quality Standards for Particulate
Matter, Chapter 5. Available on the Internet at http://www.epa.gov/ttn/ecas/regdata/RlAs/Chapter%205--
Beneflts.pdf.
4 In this analysis, the monetized benefits of reducing VOCs only reflect their effects as a PM2 5 precursor pollutant.
In this analysis, we are not quantifying any ozone-related health benefits.
7-1
-------
valuation estimates obtained in the expert elicitation study in the context of an R1A. The 14
estimates presented below derive from the application of three alternative methods:
¦ One estimate is based on the concentration-response (C-R) function developed from
the study of the American Cancer Society (ACS) cohort reported in Pope et al.
(2002), which has previously been reported as the primary estimate in recent RIAs
¦ One estimate is based on Laden et al.'s (2006) reporting of the extended Six Cities
cohort study; this study is a more recent PM epidemiological study that was used as
an alternative in the PM NAAQS RIA.
¦ The other 12 estimates are based on the results of EPA's expert elicitation study on
the PM-mortality relationship, as first reported by Industrial Economics (2006) and
interpreted for benefits analysis in EPA's final RIA for the PM NAAQS, published in
September 2006 (EPA, 2006). For that study, 12 experts (labeled A through L)
provided independent estimates of the PM-mortality C-R function. EPA practice has
been to develop independent estimates of PM-mortality estimates corresponding to
the concentration-response function provided by each of the 12 experts.
EPA believes that these updated estimates will better characterize the uncertainty
associated with using the benefit-per-ton approach to derive an estimate of total benefits. Readers
interested in the complete methodology for creating the benefit-per-ton estimates used in this
analysis may consult the Technical Support Document accompanying the final Ozone NAAQS
RIA (EPA, 2008).
To develop the estimate of the benefits of reducing emissions from this rulemaking, we
calculated the monetized benefits-per-ton of emissions reduction estimates for direct PM2sand
each PM2 5 precursor pollutant.5 In the TSD, we describe in detail how we generated the benefit-
per-ton estimates. In summary, we used a model to convert emissions of direct PM2 5 and PM2 5
precursors (i.e., SO2, NOx, and VOCs) into changes in PM25 air quality. Next, we used the
benefits model to estimate the changes in human health based on the change in PM2 5 air quality.
Finally, the monetized health benefits were divided by the emission reductions to create the
benefit per ton estimates. Even though all fine particles are assumed to have equivalent health
effects, the benefit-per-ton estimates vary because each ton of precursor reduced has a different
propensity to become PM2 5. For example, NOx has a lower benefit-per-ton estimate than direct
PM2 5 because it does not form as much PM2 5, thus the exposure would be lower, and the
monetized health benefits would be lower.
5 Emission reductions shown in the tables below are in terms of total PM, but it is estimated that 83.3% of the PM
emissions are in the PM2 5 fraction. For the purposes of this benefits analysis, all of the PM benefits shown in the
following tables are the PM25 fraction. Therefore, we do not provide any benefits estimates for the reductions of
PM other than PM2s (e.g., PM|0) that will take place as a result of this final NSPS.
7-2
-------
After generating the benefit-per-ton estimate, we then multiply this estimate by the
number of tons of each pollutant reduced to derive an overall monetary value of benefits. We
show a range of benefits estimates per pollutant (and option) rather-than a single point estimate
in order to reflect the range of estimates obtained in the expert elicitation study.6 Table 7-1
provides a general summary of the results by pollutant for the selected options, including the
emissions reductions and monetized benefits-per-ton range. Figure 7-1 provides a visual
representation of the full range of benefits estimates by pollutant at a discount rate of 3%. Tables
7-2 and 7-3 summarize the range of benefits of the selected options at discount rates of 3% and
7%, respectively. Tables 7-4 and 7-5 provide the range of benefits for all options discounted at
3% and 7% for new and modified/reconstructed units, respectively. All benefits estimates are for
the fifth year after proposal (2012). More details on the options, emissions, and emission
reductions can be found in Chapter 4 of this Rl A.
Table 7-1. General Summary of Range of Benefits Estimates for Selected Options in the
Final NSPSa
Pollutant
Emissions
Reductions
(tons)
Benefit
per Ton
(low,
3%)
Benefit
per Ton
(high,
3%)
Benefit
per Ton
(low,
7%)
Benefit
per Ton
(high,
7%)
Total Monetized
Benefits (millions
2006$ at 3%)
Total Monetized
Benefits (millions
2006$ at 7%)
Direct PM2 5
pm25
Precursor
S02
NOx
VOC
Total
1,054
$68,000
$570,000
$63,000
$520,000
$72 to
$600
$66 to $540
16,714
10,786
230
$8,000
$1,300
$210
$68,000
$11,000
$1,700
$7,400
$1,200
$190
$62,000
$9,600
$1,500
$130 to
$14 to
$.05 to
$220 to
$1,100
$110
$.38
$1,900
$120 to $1,000
$13 to $100
$.04 to $.35
$200 to $1,700
a All estimates are for the analysis year (fifth year after proposal, 2012), and are rounded to two significant figures so
numbers may not sum across columns. Emission reductions reflect the combination of selected options for both
new and reconstructed/modified sources. All benefits estimates are shown at both 3% and 7% discount rate. The
PM2 5 fraction of total PM emissions is estimated at 83.3%, and only the reduction in the PM2s fraction is
monetized in this analysis. All fine particles are assumed to have equivalent health effects, but the benefit per ton
estimates vary because each ton of precursor reduced has a different propensity to become PM2 5. The monetized
benefits incorporate the conversion from precursor emissions to ambient fine particles.
6 In the Expert Elicitation, Expert K represented the lowest estimate, and Expert E represented the highest estimate.
Therefore, the total range of benefits is presented as the range from Expert K. to Expert E.
7-3
-------
$1,200
$1,000
Epidemiology or Expert Derived FM2.5 Mortality Function
aVOC m NQX Direct PM2.5 b 902
Figure 7-1. Monetized Benefits for Selected Options for Final Petroleum Refineries NSPS
at 3% Discount Rate by PM2.5 Precursor Emitted in 2012"
a This graph shows 14 PM benefits estimates, which are treated as independent and equally probable, for each
precursor pollutant. All fine particles are assumed to have equivalent health effects, but the benefit-per-ton
estimates vary because each ton of precursor reduced has a different propensity to become PM2 5. The monetized
benefits incorporate the conversion from precursor emissions to ambient fine particles.
It is important to note that the monetized benefit-per-ton estimates used here reflect
specific geographic patterns of emissions reductions and specific air quality and benefits
modeling assumptions. Use of these $/ton values to estimate benefits associated with different
emission control programs (e.g., for reducing emissions from large stationary sources like EGUs)
may lead to higher or lower benefit estimates than if benefits were calculated based on direct air
quality modeling. Great care should be taken in applying these estimates to emission reductions
occurring in any specific location, as these are all based on national or broad regional emission
reduction programs and therefore represent average benefits-per-ton over the entire United
States. The benefits-per-ton for emission reductions in specific locations may be very different
than the national average.
7-4
-------
Table 7-2. Summary of Monetized Benefits for Selected Options at 3% Discount Rate in
2012 (millions of 2006$)
Emissions
Selected
Reduction
Total Benefit
Process Unit (new)
Option"
Pollutant6
(tons)
(millions)0
FCCU
2
PM
235
$13 -
$110
2
S02
1,993
$16 -
$140
2
NO,
859
$1.1 -
$9.1-
Small SRP
, 2
S02
42
$.34 -
$2.9
Fuel gas combustion
2
S02
524
$4.2 -
$36
Flaring gas minimization
3
S02
24
$.20 -
$1.7
-
3
NO,
9
$.01 -
$.09
3
VOC
277
$.06 -
$.46
Delayed cokers
2
so2
174
$1.4 -
$12
2 .
VOC
10
$.00 -
$.02
Process heater
2
NO,
5,237
$6.9 -
$56
Sulfur pits
2
S02
30
$.24 -
$2.0
Total
/
$43 -
$370
Process Unit
(modified/reconstructed)
FCCU
1
PM
0
$.00 -
$.00
2
S02
2,350
$19 -
$160
1
2
NOx
1,784
$2.3 -
$19
Fluid coker
2
PM
1,030
$47 -
$400
2
S02
5,893
$59 -
$490
Fluid coker
3
NO,
657
~ $.86 -
$7.0
SRP
2
> SO,
381
$3.1 -
$26
Fuel gas combustion
2
so2
4,717
$38 -
$320
Process heaters
2
NO,
2,244
$3.0 -
$24
Flaring gas minimization
3
SO,
97
$.78 -
$6.6
3
NO,
35
$.05 -
$.37
3
VOC
1,108
$.23 -
$1.8
Delayed cokers
2
S02
261
$2.1 -
$18
2
¦ VOC
15
$.00 -
$.02
Sulfur pits
2
so2
269
$2.2 -
$18
Total
$180 -
$1,500
a Refer to Chapter 4 of this R1A for more details on options.
b The PM2 5 fraction is estimated at 83.3% of the total PM emissions, and only reductions in the PM2 5 fraction is
monetized in this analysis. All fine particles are assumed to have equivalent health effects, but the benefit-per-ton
estimates vary because each ton of precursor reduced has a different propensity to become PM2 5. The monetized
benefits incorporate the conversion from precursor emissions to ambient fine particles.
c All estimates rounded to two significant figures and may not sum across columns.
7-5
-------
Table 7-3. Summary of Monetized Benefits for Selected Options at 7% Discount Rate in
2012 (millions of 2006$)
Emissions
Selected
Reduction
Total Benefit
Process Unit (new)
Option"
Pollutantb
(tons)
(millions)''
FCCU
2
PM
235
$12
$100
2
SO-,
1.993
$15
$120
2
NO,
'859
$1.0
$8.2
Small SRP
2
SO,
42
$.31
$2.6
Fuel gas combustion
2
' SO,
524
¦ $3.9
$32
Flaring gas minimization
3
SO,
16
$.12
$1.0
3
NO,
1
$.00
$.01
3*
VOC
41
$.01
$.06
Delayed cokers
2
S02
174
$1.3
$11
2
VOC
10
$.00
$.02
Process heater
2
NO,
5,237
$6.3
$50
Sulfur pits
2
SO,
30
$.22
$1.8
Total
$40
$320
Process Unit
(modified/reconstructed)
FCCU
1
PM
0
$.00
$.00
2
SO,
2,350
$17
$140
2
NO,
1,784
$2.2
$17
Fluid coker
2
PM
1.030
$43
$360
2
SO,
5,893
$54
$440
Fluid coker
3
NO,
657
$.79
$6.3
SRP
2
SO,
381
$2.8
$23
Fuel gas combustion
2
SO,
4,717
$35
$290
Process heaters
2
NO,
2.244
$2.7
$21
Flaring gas minimization
3
SO,
64
$.47
$3.9
3
NO,
4
$.01
$.04
3
VOC
165
$.03
$.3
Delayed cokers
2
SO,
261
$1.9
$16
2
VOC
15
$.00
$.02
Sulfur pits
2
SO,
269
$2.0
$17
Total
$160
$1,300
3 Refer to Chapter 4 of this RIA for more details on options.
b The PM2 5 fraction is estimated at 83.3% of the total PM emissions, and only reductions in the PM, 5 fraction is
monetized in this analysis. All fine particles are assumed to have equivalent health effects, but the beneflt-per-ton
estimates vary because each ton of precursor reduced has a different propensity to become PM, 5. The monetized
benefits incorporate the conversion from precursor emissions to ambient fine particles.
c All estimates rounded to two significant figures and may not sum across columns.
7-6
-------
Table 7-4. Estimated Range of Monetized Benefits in 2012 for All Options of New Process Units (thousands of 2006$)"
Total Benefits
Total Benefits
Total Benefits
Total Benefits
Pollutant8
Low 3%
High 3%
Low 7%.
High 7%
Process Unit
Option
($l,000/yr)
($l,000/yr)
($l,000/yr)
($l,000/yr)
FCCU
PM
1
Baseline (1.0 Ib/klb coke bum (M5B or 5F))
$—
$—
$—
$—
2
0.5 lb/klb coke'burn (M5B or 5F)C
$13,000
$110,000
$12,000
$100,000
3
0.5 lb/klb coke burn (M5)
$17,000
$140,000
$16,000
$130,000
FCCU
SO,
1
Baseline
$—
$—
$—
$—
2 25 ppmvc
$16,000
$140,000
$15,000
$120,000
FCCU
NOx
1
150 ppmv
$480
$3,900
$450
$3,500
2
80 ppmvc
$1,100
$9,100
$1,000
$8,200-
3
20 ppmv
$1,800
$15,000
$1,700
$13,000
Small SRP
SO;
1
Baseline
$—
$—
$—
$—
2
Less than 20 ltpd @ 99%c
$340
$2,900
$310
$2,600
3
All at 250 ppmv
$420
$3,500
$380
$3,200
Fuel gas
SO;
1
Baseline
$—
$—
$—
$—
combustion
2
long term limit of 60 ppmv H2SC
$4,200
$36,000
$3,900
$32,000
3
TRS limits of 160/60 ppm
$7,400
$63,000
$6,800
$57,000
Process heaters
NOx
1
80 ppmv >20 MMBtu/hr
$6,400
$51,000
$5,900
$46,000
-
2
40 ppmv >40 MMBt/hrc
$6,900
$56,000
$6,300
$50,000
3
20 ppmv, 40 MMBtu/hr
$7,700
$62,000
$7,100
$56,000
Flare gas
so2/voc
1
Baseline (no standard)
$—
$—
$—
$—
minimization
¦
2
RCA >500 lb/day S02
$120
$1,000
$110
$910
3
Option 2 + Flare minimization planc
$140
$1,200
$130
$1,100
V
4
Option 2 + No routine flaring
$140
$1,200
$130
$1,100
Delayed cokers
so,/voc •
1
Depressure to control to 15 psig
$—
$—
$—
$—
2
Depressure to control to 5 psigc
$1,400
$12,000
$1,300
$11,000
3
Depressure to control to 2 psig
$1,800
$15,000
$1,700
$14,000
(continued) ¦
-------
Table 7-4. Estimated Range of Monetized Benefits in 2012 for All Options of New Process Units (thousands of 2006$)"
(continued)
Process Unit
Pollutant8 Option
Total Benefits
Low 3%
($!,000/yr)
Total Benefits
High 3%
(S1,000/yr)
Total Benefits
Low 7%
($1,000/yr)
Total Benefits
High 7%
(Sl,000/yr)
Sulfur pits
S02 1 Do not include sulfur pits
$—
$—
$—
$—
2 Include primary sulfur pits0
$240
$2,000
$220
$1,800
3 Include primary pits and secondary tanks
$250
$2,100
$230
$1,900
° All estimates rounded to two significant figures.
b The PM2S fraction is estimated at 83.3%,of the total PM emissions, and only the reduction in the PM2j fraction is monetized in this analysis. All fine particles
are assumed to have equivalent health effects, but the benefit per ton estimates vary because each ton of precursor reduced has a different propensity to become
PM2 5. The monetized benefits incorporate the conversion from precursor emissions to ambient fine particles.
c This is the selected option.
-------
Table 7-5. Estimated Range of Monetized Benefits in 2012 for All Options of Modified/Reconstructed Process Units
(thousands of 2006$)"
Total Benefits
Total Benefits
Total Benefits
Total Benefits
Low 3%
High 3%
Low 7%
High 7%
Process Unit
Pollutantb
Option
($l,000/yr)
($l,000/yr)
($l,000/yr)
($l,000/yr)
FCCU
PM
1 Baseline (1.0 Ib/klb coke burn (M5B or 5F))C
$—
$—
$—
$—
2 0.5 Ib/klb coke burn (M5B or 5F)
$39,000
$330,000
$36,000
$300,000
3 0.5 Ib/klb coke burn (M5)
$46,000
$380,000
$42,000
$350,000
FCCU
so2
1 Baseline
$—
$—
$—
$—
2 25 ppmvc
$19,000
$160,000
$17,000
$140,000
FCCU
NO,
1 150ppmv
$1,100
$9,100
$1,000
$8,200
2 80 ppmvc
$2,300
$19,000
$2,200
$17,000
3 20 ppmv
$4,300
$34,000
$3,900
$31,000
Fluid coker
pm/so2
1 Baseline
$—
$—
$—
$—
2 1.0 lb/klb coke burn (M5B or 5F)/25 ppmvc
$110,000
$890,000
$97,000
$810,000
Fluid coker
NOx
1 Baseline
$—
$—
$—
$—
2 80 ppmvc
$540
$4,300
$500
$3,900
3 20 ppmv
$980
$7,900
$900
$7,100
Small SRP
S02
1 Baseline
$—
$—
$—
$—
2 Less than 20 Itpd @ 99%c
$3,100
$26,000
$2,800
$23,000
3 All at 250 ppmv
$3,700
$32,000
$3,400
$29,000
Fuel gas
S02
1 Baseline
$—
"$—
£—
$—
combustion
2 long term limit'of 60 ppmv H2SC
$38,000
$320,000
$35,000
$290,000
3 TRS limits of 160/60 ppm
$73,000
$620,000
$67,000
$560,000
Process heaters
NOx
1 80 ppmv >20 MMBtu/hr
$2,700
$22,000
$2,500
$20,000
2 40 ppmv >40 MMBt/hrc
$3,000
$24,000
$2,700
$21,000
3 20 ppmv. 40 MMBtu/hr
$3,300
$27,000
$3,000
$24,000
Flare gas
SCK/VOC
1 Baseline (no standard)
$—
$—
$—
$—
minimization
2 RCA >500 lb/day SO:
$470
$4,000
$430
$3,600
3 Option 2 + Flare minimization'
$550
$4,700
$510
$4,200
4 Option 2 + No routine flaring
$580
$4,900
$530
$4,400
(continued)
-------
Table 7-5. Estimated Range of Monetized Benefits in 2012 for All Options of Modified/Reconstructed Process Units
(thousands of 2006$)" (continued)
Total Benefits
Total Benefits
Total Benefits
Total Benefits
Low 3%
High 3%
Low 7%
High 7%
Process Unit
Pollutantb
Option
($l,000/yr)
($l,000/yr)
($l,000/yr)
($l,000/yr)
Delayed cokers
so:/voc
1 Depressure to control to 15 psig
$—
$—
$—
$—
2 Depressure to control to 5 psigc
S2.100
$18,000
$1,900
$16,000
3 Depressure to control to 2 psig
S2,700
$23,000
$2,500
$21,000
Sulfur pits
so.
1 Do not include sulfur pits
$—
$—
$—
$—
2 Include primary sulfur pitsc
$2,200
$18,000
$2,000
$17,000
3 Include primary pits and secondary tanks
S2.200
$19,000
$2,000
$17,000
a All estimates rounded to two significant figures.
^ The PM2 5 fraction is estimated at 83.3% of the total PM emissions, and only the reduction in the PM2 5 fraction is monetized in this analysis. All fine particles
are assumed to have equivalent health effects, but the benefit per ton estimates vary because each ton of precursor reduced has a different propensity to become
5- The monetized benefits incorporate the conversion from precursor emissions to ambient fine particles.
c This is the selected option.
-------
7.2 Characterization of Uncertainty in the Benefits Estimates
In any complex analysis, there are likely to be many sources of uncertainty. Many inputs
are used to derive the final estimate of economic benefits, including emission inventories, air
quality models (with their associated parameters and inputs), epidemiological estimates of
concentration-response (C-R) functions, estimates of values, population estimates, income
estimates, and estimates of the future state of the world (i.e., regulations, technology, and human
behavior). For some parameters or inputs it may be possible to provide a statistical representation
of the underlying uncertainty distribution. For other parameters or inputs, the necessary
information is not available.
The annual benefit estimates presented in this analysis are also inherently variable due to
the processes that govern pollutant emissions and ambient air quality in a given year. Factors
such as hours of equipment use and weather are constantly variable, regardless of our ability to
measure them accurately. As discussed in the PM25 NAAQS RIA (Table 5.5), there are a variety
of uncertainties associated with these PM benefits. Therefore, the estimates of annual benefits
should be viewed as representative of the magnitude of benefits expected, rather than the actual
benefits that would occur every year.
Above we present the estimates of the total benefits, based on our interpretation of the
best available scientific literature and methods and supported by the SAB-HES and the NAS
(NRC, 2002). The benefits estimates are subject to a number of assumptions and uncertainties.
For example, for key assumptions underlying the estimates for premature mortality, which
typically account for at least 90% of the total benefits, we were able to quantify include the
following:
1. Inhalation of fine particles is causally associated with premature death at
concentrations near those experienced by most Americans on a daily basis. Although
biological mechanisms for this effect have not been established definitively yet, the
weight of the available epidemiological evidence supports an assumption of causality.
2. All fine particles, regardless of their chemical composition, are equally potent in
causing premature mortality. This is an important assumption, because PM produced
via transported precursors emitted from EGUs may differ significantly from direct
PM released from diesel engines and other industrial sources, but no clear scientific
grounds exist for supporting differential effects estimates by particle type.
3. The impact function for fine particles is approximately linear within the range of
ambient concentrations under consideration. Thus, the estimates include health
benefits from reducing fine particles in areas with varied concentrations of PM,
7-11
-------
including both regions that are in attainment with fine particle standard and those that
do not meet the standard.
4. The forecasts for future emissions and associated air quality modeling are valid.
Although recognizing the difficulties, assumptions, and inherent uncertainties in the
overall enterprise, these analyses are based on peer-reviewed scientific literature and
up-to-date assessment tools, and we believe the results are highly useful in assessing
this rule.
5. Benefits estimated here reflect the application of a national dollar benefit-per-ton
estimate of the benefits of reducing directly emitted fine particulates from point
sources. Because they are based on national-level analysis, the benefit-per-ton
estimates used here do not reflect local variability in meteorology, exposure, baselinev
health incidence rates, or other local factors that might lead to an over-estimate or
under-estimate of the actual benefits of controlling directly emitted fine particulates.
This RIA does not include the type of detailed uncertainty assessment found in the PM
NAAQS RIA because we lack the necessary air quality input and monitoring data to run the
benefits model. Moreover, it was not possible to develop benefit-per-ton metrics and associated
estimates of uncertainty using the benefits estimates from the PM RIA because of the significant
differences between the sources affected in that rule and those regulated here. However, the
results of the Monte Carlo analyses of the health and welfare benefits presented in Chapter 5 of
the PM RIA can provide some evidence of the uncertainty surrounding the benefits results
presented in this analysis.
7.3 Updating the Benefits Data Underlying the Benefit-per-Ton Estimates
As described above, the estimates provided in Tables 7-1 through 7-5 are derived through
a benefits transfer technique that adapts monetized benefits from reductions in PM2 5 precursor
pollutants that were estimated for the Ozone RIA utilizing nationally distributed emissions
reductions. EPA is currently in the process of generating localized benefit-per-ton estimates to
better account for the spatial heterogeneity of benefits. EPA believes that these localized
estimates may better represent the actual benefits than estimates that use national averages.
7.4 Comparison of Benefits and Costs
EPA estimates the range of annualized benefits of this rulemaking to be a combined $220
million to $1.9 billion ($2006) for new and reconstructed/modified sources at a 3% discount rate
and annualized costs calculated at a 7% interest rate as mentioned in Chapter 4 of this RIA for
these sources to be $31 million ($2006) in the fifth year after proposal (2012). Thus, net benefits
are $190 million to $1.8 billion in the fifth year after proposal at a 3% discount rate for the
benefits. Figure 7-2 shows the full range of net benefits estimates (i.e., annual benefits in 2012
7-12
-------
minus annualized costs) utilizing the 14 different PM2 5 mortality functions at the 3% discount
rate. EPA believes that the benefits are likely to exceed the costs by a substantial margin under
this rulemaking even when taking into account uncertainties in the cost and benefit estimates.
$2,000 -
$1,500 |
o
g -$500 -
-$1,000 J
-$1,500 J
-$2,000 !
Various Cbmbinations of Cbsts and Benefits Estimates
Figure 7-2. Range of Estimated Net Benefits for Selected Options for Final Petroleum
Refineries NSPSa
a Net Benefits are quantified in terms of PM2 5 benefits at a 3% discount rate for the fifth year after proposal. This
graph shows 14 benefits estimates combined with the cost estimate. All combinations are treated as independent
and equally probable. All fine particles are assumed to have equivalent health effects, but the benefit-per-ton
estimates vary because each ton of precursor reduced has a different propensity to become PM2 5. The monetized
benefits incorporate the conversion from precursor emissions to ambient fine particles.
7.5 References
Ayyb, 2002
Guo et al., 1999.
Ibald-Mulli et al., 2001.
Industrial Economics, Inc., 2006. Expanded Expert Judgment Assessment of the Concentration-
Response Relationship Between PM2.5 Exposure and Mortality. Prepared for the U.S.
EPA, Office of Air Quality Planning and Standards, September.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air
Pollution Regulations. Washington, DC: The National Academies Press.
7-13
-------
Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. Reduction in Fine Particulate Air
Pollution and Mortality. American Journal of Respiratory and Critical Care Medicine.
173: 667-672.
Pope, C.A., III, R.T. Burnett, MJ. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston.
2002. "Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine
Particulate Air Pollution." Journal of the American Medical Association 287:1132-1141.
U.S. Environmental Protection Agency. 2006. Regulatory Impact Analysis, 2006 National
Ambient Air Quality Standards for Particulate Matter, Chapter 5. Available at
.
7-14
-------
APPENDIX C
OVERVIEW OF ECONOMIC MODEL EQUATIONS
We illustrate our approach for addressing conceptual questions of market-level
impacts using a numerical simulation model. Our method involves specifying a set of
nonlinear supply and demand relationships for the aiffected markets, simplifying the
equations by transforming them into a set of linear equations, and then solving the
equilibrium system of equations (see Fullerton and Metcalfe [2002] for an example).
C.l Discussion and Specification of Model Equations
First, we consider the formal definition of the elasticity of supply with respect to
changes in own price:
dQi / Qs
ts = 5 5 (C.l)
dp/p
Next, we can use "hat" notation to transform Eq. (D.l) to proportional changes and
rearrange terms:
Qs = *sP (Cla)
where
Qs = percentage change in the quantity of market supply,
es = market elasticity of supply, and
p = percentage change in market price.
As Fullerton and Metcalfe (2002) note, we have taken the elasticity definition and turned
it into a linear behavioral equation for our market.
To introduce the direct impact of the regulatory program, we assume the per-unit
cost associated with the regulatory program (c)1 leads to a proportional shift in the
marginal cost of production. Under the assumption of perfect competition (price equals
marginal cost), we can approximate this shift at the initial equilibrium point as follows:
'The per-unit costs (c) are computed by dividing the total annualized costs reported in by the baseline
consumption.
1
-------
MC =
c c
(C.lb)
MC0 p0
The with-regulation supply equation can now be written as
Qs=es(p-MC).
(C.lc)
Next, we can specify a demand equation as follows:
Qd = v/p
(C.2)
where
Qd - percentage change in the quantity of market demand,
r\d = market elasticity of demand, and
p = percentage change in market price.
Finally, we specify the market equilibrium conditions in the affected markets. In
response to the exogenous increase in production costs, producer and consumer behaviors
are represented in Eq. (C.la) and Eq. (C.2), and the new equilibrium satisfies the
condition that the change in supply equals the change in demand:
We now have three linear equations in three unknowns (p , Qd, and Qs ), and we
can solve for the proportional price change in terms of the elasticity parameters (es and
r|
-------
x MC.
ts-^d
(C.5)
Given this solution, we can solve for the proportional change in market quantity using
Eq.(C.2).
C.2 Consumer and Producer.Welfare Calculations
The change in consumer surplus in the affected markets can be estimated using
the following linear approximation method:
ACS = - [Qi x Ap] + [0.5 x AQx Ap].
As shown, higher market prices and reduced consumption lead to welfare losses for
consumers. A geometric representation of this calculation is illustrated in Figure C-l.
$
(C.6)
Price
Increase
{
Pi
Po
Si: With Regulation
CO
^ Unit Cost Increase
S0: Without Regulation
f
c
d
Qi Qo Output
A consumer surplus = -[fghd + dhc]
A producer surplus = [fghd - aehb] - bdc
A total surplus = -[aehb +>dhc + bdc]
Figure C-l. Welfare Calculations
For affected supply, the change in producer surplus can be estimated with the
following equation:
3
-------
APS = [Q, x Ap] - [Q, x C] - [0.5 x AQ x (4, - c)). (C.7)
Increased regulatory costs and output declines have a negative effect on producer surplus,
because the net price change (Ap - c) is negative. However, these losses are mitigated, to
some degree, as a result of higher market prices. A geometric representation of this
calculation is illustrated in Figure C-l.
4
------- |