Final Technical Support Document (TSD)
for the Cross-State Air Pollution Rule for the 2008 Ozone NAAQS

Docket ID No. EPA-HQ-OAR-2015-0500

Assessment of Non-EGU NOx Emission Controls, Cost of
Controls, and Time for Compliance Final TSD

U.S. Environmental Protection Agency
Office of Air and Radiation
August 2016

1


-------
1 Introduction/Purpose

The purpose of this Technical Support Document (TSD) is to discuss the currently available
information on emissions and control measures for sources of NOx other than electric
generating units (EGUs). This information provides more detail about why EGUs are the
focus of the final rulemaking, namely the uncertainty that exists regarding whether
significant aggregate NOx mitigation is achievable from non-EGU point sources by the 2017
ozone season, and the fact that the limited available information points to an apparent
scarcity of non-EGU reductions that could be accomplished in this timeframe.
Notwithstanding these conclusions as regards the 2017 ozone season, the EPA continues to
assess the role of NOx emissions from non-EGU sources to downwind nonattainment
problems.

This TSD begins by briefly discussing the non-EGU emissions inventories used in the
proposed and final Cross-State Air Pollution Rule (CSAPR) Update analyses, both for the
2011 base year and 2017 future baseline assessed for this rule. The TSD then presents an
evaluation of whether non-EGU emissions can be reduced in a cost-effective manner for
particular categories. Then, it assesses the available NOx emission reductions from such
categories and presents the category-by-category emissions reduction potential. This
assessment considers and presents the annualized costs per ton of these reductions, with a
focus on technologies that achieve cost-effective reductions within a range of costs similar
to that evaluated for EGUs. The TSD then presents estimates of the time required to install
and implement the control measures, both for comparison to the 2017 compliance
timeframe, and for discussion of installation time should such measures be required in the
future. It should be noted that no changes to these data or estimates have been made for
this final TSD compared to the draft version of this TSD provided in the docket for the
proposed rule. Finally, the TSD presents a summary of comments received on the proposed
rule TSD, along with responses as appropriate.

For the reasons stated in the preamble, the data and discussion in this TSD are intended to
focus on the eastern states that are included in the CSAPR Update rule. Information
inclusive of western states1 is presented where available and appropriate.

1 For the purpose of this action, the western United States (or the West] consists of the 11 western contiguous states of
Arizona, California, Colorado, Idaho, Montana, New Mexico, Nevada, Oregon, Utah, Washington, and Wyoming, and the
eastern U.S. (or East] consists of the remaining states in the contiguous U.S.

2


-------
2 Background

In this section we present annual and ozone-season NOx emission inventory totals and the
relative percentages for non-EGU source categories statewide and/or nationally. This
information is summary in nature and is not meant to replace other, more detailed
information available from the EPA, such as the EPA's 2011v6.2 Emissions Modeling
Platform TSD2 as well as the Notice of Data Availability3 (NODA) and Regulatory Impact
Analysis4 (RIA) for the proposed and final rule.

Table 1 lists 2011 and 2017 projected NOx emissions by sector, in summary form, for the
48 contiguous states of the United States (CONUS).

Table 1: 2011 Base Year and 2017 Projected NOx Emissions by
	Sector (tons), for the 48 CONUS	

Sector

2011 NOx,
annual

2017 NOx,
annual

2011 NOx, ozone
season

2017 NOx, ozone
season

EGU-point

2,000,000

1,500,000

942,000

689,000

NonEGU-point

1,200,000

1,200,000

515,000

502,000

Point oil and gas

500,000

410,000

213,000

172,000

Wild and prescribed fires

330,000

330,000

165,000

165,000

Nonpoint oil and gas

650,000

690,000

275,000

293,000

Residential wood
combustion

34,000

35,000

3,000

3,000

Other nonpoint

760,000

730,000

204,000

211,000

Nonroad

1,600,000

1,100,000

825,000

582,000

Onroad

5,700,000

3,200,000

2,417,000

1,329,000

C3 commercial marine
vessel (CMV)

130,000

130,000

58,000

58,000

Locomotive and C1/C2
CMV

1,100,000

910,000

451,000

384,000

Biogenics

1,000,000

1,000,000

630,000

630,000

TOTAL

15,000,000

11,200,000

6,698,000

5,018,000

It is clear from Table 1 that NOx emissions are projected to remain constant or decrease for
most sectors in the 48 states between 2011 and 2017, and this is true whether examining
annual or ozone season (OS) tons. Emissions from the non-EGU point source sector and the
other nonpoint source sector are not projected to change significantly, while emissions

2	Technical Support Document (TSD], Preparation of Emissions Inventories for the Version 6.2,2011 Emissions Modeling
Platform, August 2015, available at: https://www.epa.gov/air-emissions-modeling/2011-version-62-technical-support-
document

3	Notice of Availability of the Environmental Protection Agency's Updated Ozone Transport Modeling Data for the 2008
Ozone National Ambient Air Quality Standard (NAAQS]. The official version is available in the docket for this rulemaking.

4	Regulatory Impact Analysis for the Proposed Cross-State Air Pollution Rule (CSAPR] for the 2008 Ozone National
Ambient Air Quality Standards (NAAQS] and Regulatory Impact Analysis for the Cross-State Air Pollution Rule (CSAPR]
Update for the 2008 Ozone National Ambient Air Quality Standards (NAAQS]. The official versions are available in the
docket for this rulemaking.

3


-------
from the nonpoint oil and gas source sector are projected to grow (approximately
6%)during this time period. Based on the values in Table 1, Figures 1 and 2 show the
relative contributions of the various sectors to overall NOx emissions (left panel) in the
CONUS and for the non-EGU sectors (right panel) for 2011 and 2017, respectively.

Figure 1:2011 NOx emissions by sector, with further non-EGU

breakout (48 states)

2011 v6.2 Emissions - NOx by	2011 v6.2 Emissions -

Sector (Total 6.7 million OS tons)	NOx in Non-EGU

Sector (further detail)

60%

EGU

Non-EGU total
Fires

Residential wood

Mobile

Biogenics

-

25%

IE

42%

Non-EGU
point

Point oil &

gas

Nonpoint oil

& gas

Other
nonpoint

Figure 2: Projected 2017 NOx emissions by sector, with further

non-EGU breakout (48 states)

2017 Projections - NOx By Sector
(TOTAL 5.0 million OS TONS)

2017 Projections -
NOx in Non-EGU
Sector (further detail)

¦	EGU

¦	Non-EGU total

¦	Fires

¦	Residential wood
Mobile
Biogenics

Non-EGU
point

¦ Point oil &
gas

¦	Nonpoint oil
& gas

¦	Other
nonpoint

4


-------
Figure 1 depicts the CONUS total ozone season NOx emissions of 6,698,000 tons in 2011
and Figure 2 depicts the CONUS total ozone season NOx emissions of 5,018,000 tons in
2017. In both 2011 and 2017, the mobile source sector has the largest NOx emissions.5
Substantial reductions in mobile source NOx emissions are projected to occur by 2017.
Mobile source emissions are projected to decrease because of sector-specific standards
related to fuels, fuel economy, pollution controls, and repair and replacement of the
existing fleet. Because these reductions are already expected to occur, mobile source
emission reductions are not included in this analysis of non-EGU emission reductions
achievable by the 2017 ozone season.

For the purposes of preliminary analysis in this TSD, "non-EGU total" refers to four
separate categories of sources: non-EGU point, point oil and gas, nonpoint oil and gas, and
other nonpoint (and does not include mobile sources). The oil and gas point and nonpoint
sources are separated from the remaining non-EGU point and nonpoint sources due to the
magnitude of their contribution to the inventory and other aspects related to the inventory
development, emissions modeling, and future year projections for that industry. The point
oil and gas sources are also separated out from the other non-EGU point sources according
to the North American Industry Classification System (NAICS) code specified for the
various sources. Note that point oil and gas sources include a variety of types of processes,
and there is overlap with the processes included in the rest of the non-EGU point inventory.
More information on the emissions sectors is available in the 2011v6.2 Emissions Modeling
Platform TSD.

Comparing the proportions of the total inventory for non-EGUs (Figures 1 and 2), it
becomes clear that, although they are decreasing in the absolute sense, non-EGU NOx
emissions are becoming a larger share of overall ozone-season NOx emissions (16% in
2011 compared with 21% in 2017).

Table 2 compares statewide projected total anthropogenic NOx emissions (inclusive of all
sectors listed in Table 1 with the exception of fires and biogenics) for the 2017 ozone
season to non-EGU NOx emissions for the 2017 ozone season for each of the 48 contiguous
United States . Totals are given for the 48 contiguous United States (the 37 eastern states
plus the District of Columbia that are addressed in the rule are highlighted below in blue).
Non-EGU sources in this table are broken down into two groups (non-EGU point sources,
including point oil & gas sources, and other nonpoint and nonpoint oil & gas sources).

5 The mobile source sector comprises multiple different types of sources (onroad cars & trucks, boats, ships, trains,
construction equipment, mining equipment, tractors, etc.].

5


-------
Table 2: Projected Total Anthropogenic Ozone-Season NOx
Emissions vs. Projected Non-EGU Source Group NOx Emissions,
		2017 Projections, Tons6 		

State

Total

Anthropogenic

Non-EGU
Point + Oil &
Gas Point

%

Anthro

Oil & Gas
Nonpoint+ Other
Nonpoint

%

Anthro

Oil & Gas
Point + Oil
& Gas
Nonpoint

%

Anthro

Alabama

88,805

22,187

25

7,952

9

7,442

8

Arizona

71,906

5,015

7

2,310

3

612

1

Arkansas

69,737

13,400

19

5,308

8

9,164

13

California

236,322

29,342

12

20,220

9

3,105

1

Colorado

90,756

19,594

22

16,899

19

27,284

30

Connecticut

17,672

1,105

6

2,626

15

98

1

Delaware

7,786

628

8

615

8

0

0

District of
Columbia

2,252

212

9

312

14

0

0

Florida

177,514

16,293

9

7,543

4

1,112

1

Georgia

103,536

18,8.16

18

4,559

4

1,495

1

Idaho

27,893

3,752

13

1,989

7

503

2

Illinois

148,178

24,668

17

15,409

10

9,424

6

Indiana

139,133

27,222

20

6,864

5

5,931

4

Iowa

70,467

7,888

11

3,861

5

153

0

Kansas

79,939

6,968

9

12,619

16

10,697

13

Kentucky

106,830

11,456

11

11,905

11

12,251

11

Louisiana

173,330

45,506

26

30,160

17

31,503

18

Maine

17,576

4,639

26

809

5

26

0

Maryland

46,029

6,213

13

3,508

8

522

1

Massachusetts

35,369

4,144

12

4,807

14

105

0

Michigan

131,486

21,867

17

12,245

9

9,398

7

Minnesota

89,328

15,541

17

6,414

7

46

0

Mississippi

54,832

11,684

21

2,122

4

6,557

12

Missouri

101,035

9,238

9

3,594

4

122

0

Montana

38,504

2,948

8

3,630

9

3,390

9

Nebraska

70,005

3,884

6

1,163

2

467

1

Nevada

28,192

4,018

14

1,003

4

115

0

New Hampshire

8,932

680

8

1,028

12

0

0

New Jersey

52,743

4,544

9

5,506

10

173

0

New Mexico

65,263

10,559

16

19,940

31

27,759

43

New York

109,910

13,738

12

14,624

13

904

1

North Carolina

98,064

15,711

16

3,657

4

1,203

1

North Dakota

74,118

4,047

5

18,125

24

19,185

26

Ohio

160,110

21,280

13

11,617

7

2,906

2

Oklahoma

131,763

32,203

24

33,178

25

51,257

39

Oregon

40,507

6,130

15

4,348

11

365

1

6 EGUs are not provided a separate breakout in Table 2 since state-level emissions are presented in the Preparation of
Emissions Inventories for the Version 6.3,2011 Emissions Modeling Platform TSD and other TSDs for the proposed and
final rules.

6


-------












Oil & Gas







Non-EGU



Oil & Gas



Point + Oil





Total

Point + Oil &

%

Nonpoint+ Other

%

& Gas

%

State

Anthropogenic

Gas Point

Anthro

Nonpoint

Anthro

Nonpoint

Anthro

Pennsylvania

174,664

23,735

14

33,508

19

26,713

15

Rhode; Island

5,845

544

9

1,370

23

12

0

South Carolina

55,897

10,144

18

3,980

7

348

1

South Dakota

22,192

1,24.1

6

432

2

75

0

Tennessee

85,759

13,494

16

5,846

7

1,922

7

Texas

467,245

95,671

20

115,180

25

145,285

31

Tribal Data

26,717

3,799

14

0

0

3,700

14

Utah

66,486

8,004

12

9,781

15

9,349

14

Vermont

5,473

163

3

937

17

0

0

Virginia

87,754

14,039

16

7,318

8

4,775

5

Washington

75,833

8,666

11

1,150

2

164

0

West Virginia

64,839

9,678

15

12,642

19

16,723

26

Wisconsin

75,047

11,181

15

5,351

7

178

0

Wyoming

68,864

26,488

38

4,018

6

10,905

16

Eastern States

3,411,193

545,649

16

418,692

12

378,171

11

US Total

4,248,436

673,964

16

503,980

12

465,421

11

Table 2 indicates that, in the projected 2017 inventory, non-EGU sources comprising non-
EGU point and point oil and gas sources are estimated to make up 16% of anthropogenic
NOx emissions in the 48 contiguous United States. In individual states, the percentage of
anthropogenic emissions contributed by these two non-EGUs categories range from 3% to
26% (eastern states) and from 7% to 38% (western states).

We also note that in the projected 2017 inventory, non-EGU sources comprising nonpoint
oil & gas and other nonpoint sources are estimated to make up 12% of anthropogenic NOx
emissions in the entire continental U.S. In individual states, the percentage of
anthropogenic emissions contributed by these non-EGUs ranges from 2% to 25% (eastern
states) and from 4% to 31% (western states).

The EPA's preliminary analysis indicates that NOx emissions from oil and gas sources
(inclusive of emissions from the point oil and gas and nonpoint oil and gas sectors)
comprise an average of 11% of the total ozone season NOx emissions inventory. For some
states, this percentage increases up to 43%, with oil and gas emissions exceeding non-EGU
point totals in a number of states. The key sources of NOx emissions in the oil and gas
sector are from the combustion of fossil fuel (primarily drilling rigs, internal combustion
(IC) engines and pipeline compressors) and flares. Please refer to the EPA's 2011v6.2
Emissions Modeling Platform TSD for more information on emissions from these sectors.

7


-------
3 Preliminary Analysis

For the purposes of the proposed rule, the EPA performed a preliminary analysis to
characterize whether there are non-EGU source groups with a substantial amount of
available cost-effective NOx reductions achievable by the 2017 ozone season. The EPA
received no comments that would substantively change this analysis, therefore there was
no need to repeat this preliminary analysis for the final CSAPR Update rule.

3.1 Methodology

The EPA's preliminary analysis of potential non-EGU NOx emission reductions was
performed using the Control Strategy Tool (CoST). CoST is the software tool the EPA uses
to estimate the emission reductions and costs associated with future-year control
strategies, and then to generate emission inventories that result from the control strategies
applied. CoST tracks information about control measures, their costs, and the types of
emissions sources to which they apply. The purpose of CoST is to support national- and
regional-scale multi-pollutant analyses, primarily for Regulatory Impact Analyses (RIAs) of
the National Ambient Air Quality Standards (NAAQS). CoST is also a component of the
Emissions Modeling Framework (EMF) that was used to generate the 2017 non-EGU
emissions presented above and in the Emissions Modeling Platform TSD for the proposed
CSAPR Update rule. Further discussion and documentation of CoST is available on the
EPA's website at http://www.epa.gov/ttnecasl/cost.htm.

Appendices to this TSD discuss recommendations for updates to CoST, including
corrections for inapplicable controls, sources already controlled by state rules, sources
with permit emissions limits or that have clearly identified controls in place, and sources
subject to future NOx emission limits. Appendix A discusses contractor RTI International's
work to review estimates for lean burn internal combustion (IC) engines, glass
manufacturing, ammonia reformers, and gas turbines.7 Appendix B discusses contractor
SRA International's work on a variety of other categories including many of the others
evaluated in this TSD.8

EPA has prepared a set of data called the Control Measure Data Base (CMDB) that is used as
an important input to CoST. This data includes all control measures utilized by the tool for
control strategy analysis. It should be noted that most of the NOx measures included in this
report are currently in the Control Measure Data Base used by CoST, and generally do not
reflect the updates suggested in these contractor reports. Obstacles to full incorporation of
the recommended changes include availability of accurate costs for these measures, and to
have cost equations rather than average cost/ton to estimate costs. Control efficiencies are
readily available for measures, but costs, particularly those that can be estimated using
equations that consider source size or capacity, often are not. In addition, the Pennsylvania

7	"Update of NOx Control Measure Data in the CoST Control Measure Database for Four Industrial Source Categories:
Ammonia Reformers, NonEGU Combustion Turbines, Glass Manufacturing, and Lean Burn Reciprocating Internal
Combustion Engines," Revised Draft Report, RTI International, 2014.

8	"Review of CoST Model Emission Reduction Estimates," SRA International, 2014; "Summary of State NOx Regulations for
Selected Stationary Sources," SRA International, 2014.

8


-------
Department of Environmental Protection's Additional RACT Requirements for Major
Sources of NOx and VOCs rule,9 which was recently finalized, is not included in these
contractor reports. The EPA will consider whether or not to incorporate these
recommendations for changes or additions to the NOx controls for non-EGUs to support
NOx control efforts for future rules and other efforts. The EPA will also consider updates to
reflect state emission control requirements (i.e., Pennsylvania's RACT rule). Nonetheless,
the information from these reports helped inform our assessment in terms of uncertainty
surrounding non-EGU emission reduction potential. Further details on the CMDB can be
found on the CoST web site shown above.

For the purpose of identifying a list of non-EGU NOx source groups with controls available,
the EPA ran CoST for non-EGU point sources for the 37 eastern U.S. with NOx emissions of
greater than 25 tons/year in 2017. The analysis using CoST was a basis for the review of
NOx control measures for non-EGUs undertaken by two different contractors for EPA.
Through a contractual agreement with EPA, SRA International and RTI International
provided reports within which CoST examined a number of source categories of non-EGUs
with annualized control costs up to $10,000 per ton (in 2011 dollars). These reports are
included in the Appendices of this TSD. CoST selected particular control technologies based
on application of a least-cost criterion for control measures applied as part of the control
strategy. Other NOx control measures are available for some of these categories, but on
average, annualized costs for these measures were at higher cost.

3.2 Uncertainties and Limitations

The EPA acknowledges several important limitations of the non-EGU cost analysis included
in this TSD, which include the following:

Boundary of the cost analysis: In this cost analysis we include only the impacts to the
regulated industry, such as the costs for purchase, installation, operation, and maintenance
of control equipment over the lifetime of the equipment. Recordkeeping, reporting, testing
and monitoring costs are not included. Additional profit or income may be generated by
industries supplying the regulated industry, especially for control equipment
manufacturers, distributors, or service providers. These types of secondary impacts are not
included in this cost analysis.

Cost and effectiveness of control measures: Our application of control measures reflect
nationwide average retrofit factors and equipment life. We do not account for regional or
local variation in capital and annual cost items such as energy, labor, materials, and others.
Our estimates of control measure costs may over- or under-estimate the costs depending
on how the difficulty of actual retrofitting and equipment life compares with our control
assumptions. In addition, our estimates of control efficiencies for control measures
included in our analysis assume that the control devices are properly installed and
maintained. There is also variability in scale of application that is difficult to reflect for
small area sources of emissions.

9 Available at: http://www.pabulletin.com/secure/data/vol46/46-17/694.html

9


-------
Discount (interest) rate: Because we obtain control cost data from many sources, we are
not always able to obtain consistent data across original data sources. If disaggregated
control cost data are not available (i.e., where capital, equipment life value, and operation
and maintenance [O&M] costs are not shown separately), the EPA assumes that the
estimated control costs are annualized using a 7 percent discount rate, which is the
discount (interest) rate used in accordance with OMB guidance in Circular A-4. In general,
we have some disaggregated data available for non-EGU point source controls. In addition,
while these interest rates are consistent with OMB guidance, the actual interest rates may
vary regionally or locally.

Accuracy of control costs: We estimate that there is an accuracy range of +/- 30 percent for
non-EGU point source control costs. This level of accuracy is described in the EPA Air
Pollution Control Cost Manual, which is a basis for the estimation of non-EGU control cost
estimates included in this TSD. This level of accuracy is consistent with either the budget or
bid/tender level of cost estimation as defined by the American Association for Cost
Engineering (AACE) International. In addition, the accuracy of costs is also influenced by
the availability of data underlying the cost estimates for individual control measures. For
some control measures, we recognize that there is limited data available to generate robust
cost estimates. This is reflected in the derivation of costs for some of the non-EGU NOx
control measures discussed in Appendix A for this TSD.

3.3 CoSTResults

The results of the CoST analysis are displayed in Table 3. In Table 3, we display the source
groups selected by CoST, the Source Classification Codes (SCCs) included in those groups10,
the least-cost control technology for a given source group (selected by CoST), the current
estimate (in dollars per ton, using 2011 dollars) of the annualized cost per ton NOx reduced
of the control technology, the current estimate of the time necessary to install the selected
control technology (not including permitting time), the estimated ozone season emissions
in the East from the non-EGU source group in 2017 in the absence of the installation of the
selected controls, and the estimated potential ozone season reductions in the East from the
non-EGU source group in 2017 assuming the CoST selected controls could be fully installed
and operational prior to the 2017 ozone season (which as discussed in more detail later, is
not the case for many of the categories examined). Note that CoST does not account for
installation time or time required for the permitting process. Instead it provides
information on the control measures applicable to sources in the inventory, along with the
cost of installation and operation and maintenance of the selected measures.

10 The CoST results do not indicate applicability of the recommended control technology to all sources in the source group
but only to the specific SCCs for which control technologies are applicable. For example, for the cement kilns source
group, Biosolid Injection Technology (BSI] is applicable only for the types of cement kilns covered by the listed SCCs.

10


-------
Table 3: CoST Results: Non-EGU Source Groups with NOx Reductions

Non-EGU

SCCs

Control

Current

Time to install11

201714

2017

Source Group



Technology

estimate

12 (excluding

NOx

Potential





Recommen

of NOx

permitting,

Emissio

Reductions





ded by

$/ton,

reporting

ns (37

15 (37 States





CoST

CoST

preparation,

States +

+ DC), OS







(2011 $)

programmatic

DC), OS

tons, CoST









and

tons,











administrative

CoST











considerations13

1





Cement Kilns

30500622 (preheater kiln], 30500623

Biosolid

$410

Uncertain

24,760

4,207



(preheater/precalciner]; 39000201

Injection











(kiln/dryer]; 39000288 (kiln in process

Technology











coal]

(BSI]









11	Time to install is not an output of CoST, but are rather estimates determined by the EPA based on research from a variety of sources. See, "Typical Installation
Timelines for NOx Emissions Control Technologies on Industrial Sources," Institute of Clean Air Companies, December 2006 (all sources except Cement Kilns and RICE
(Reciprocating Internal Combustion Engines]], "Cement Kilns Technical Support Document for the NOx FIP," EPA, January 2001 (cement kilns], and "Availability and
Limitations of NOx Emission Control Resources for Natural Gas-Fired Reciprocating Engine Prime Movers Used in the Interstate Natural Gas Transmission Industry,"
Innovative Environmental Solutions Inc., July 2014 (prepared for the INGAA Foundation].

12	In general, for control retrofits to non-EGU sectors, it appears that the full sector-wide compliance time is uncertain, but is longer than the installation time shown
above for a typical unit. We have insufficient information on capacity and experience within the OEM suppliers and major engineering firms supply chain to offer
conclusions on their availability to execute the project work for non-EGU sectors.

13	Non-EGUs of any type - boiler or turbine - that are not currently required to monitor and report in accordance with 40 CFR Part 75 and/or not currently participating
in the CSAPR program will require additional time relative to EGUs that are currently equipped with Part 75 monitoring and reporting and/or participating in the
current CSAPR program. Installation of NOx monitors for the reporting of NOx mass requires the construction of platforms, CEM shelters, procurement of equipment,
certification testing, and electronic data reporting programming of a data handling system. These added timing considerations for infrastructure on the non-EGU
sources combined with the additional programmatic adoption measures necessary make installation of controls by the 2017 timeframe established in this rule less likely
and more uncertain for industrial sources.

14	Emissions and potential reductions for Gas Turbines ($163/ton grouping]. Cement Kiln/Dryer (Bituminous Coal] ($942/ton grouping]. Coal Cleaning - Thermal Dryer
(2], Spreader Strokers, Petroleum Refinery Process Heaters, Incinerators, Boilers & Process Heaters, Gas-Fired Process Heaters, Coal Boilers, By-Product Coke
Manufacturing, ICI Boilers - Residual Oil, Ammonia Production, Glass Manufacturing, ICI Boilers, Iron & Steel - In-Process Combustion - Bituminous Coal, Industrial
Processes Miscellaneous, Catalytic Cracking, Process Heaters, & Coke Ovens, Petroleum Refinery Gas-Fired Process Heaters, Glass Manufacturing - Pressed, Glass
Manufacturing - Container, Petroleum Refinery Gas-Fired Process Heaters, and RICE source groups were calculated for 2018, however they are likely to be virtually
identical to projections for 2017. Non-EGU source groups with projected aggregate 2017 NOx emissions below 100 OS tons are excluded from this table.

15	Potential reductions assume fully implemented controls by the start of the 2017 ozone season.

11


-------
Cement Mfg
(dry)

30500606 Industrial Processes, Mineral
Products, Cement Manufacturing (Dry
Process), Kilns

Selective Non-
Catalytic
Reduction
(SNCR]

$1,255

42-51 weeks

13,006

6,501

Cement Mfg
(wet)

30500706 Industrial processes, mineral
products. Cement Manufacturing (Wet
Process], Kilns

Mid-Kiln Firing

$73

5-7 months

7,971

2,287

Coal Cleaning -
Thermal Dryer

(1)

30502508 Construction Sand & Gravel,
Dryer; 30501001 Industrial Processes,
Mineral Products, Coal Mining, Cleaning,
and Material Handling, Fluidized Bed
Reactor

Low NOx Burner
(LNB)

$1,125

6-8 months

503

165

Coal Cleaning -
Thermal Dryer
(2)

30501001 Industrial Processes, Mineral
Products, Coal Mining, Cleaning, and
Material Handling, Fluidized Bed Reactor

Low NOx Burner
(LNB)

$1,640

6-8 months

154

63

Cement
Kiln/Dryer
(Bituminous
Coal)

39000201 Industrial Processes, In-process
Fuel Use, Bituminous Coal, Cement
Kiln/Dryer (Bituminous Coal)

SNCR

$942

42-51 weeks

520

260

Iron and Steel
Mills -
Reheating

30300934 (303015] Primary Metal
Production: Steel; 30300933

Low NOx Burner
(LNB] & Flue Gas
Recirculation
(FGR]

$620

6-8 months

1,064

664

Steel

Production

30490033 Industrial Processes,
Secondary Metal Production, Fuel Fired
Equipment, Natural Gas: Furnaces;
30400704 Industrial Processes,
Secondary Metal Production, Steel
Foundries, Heat Treating Furnace

Low NOx Burner
(LNB)

$928

6-8 months

281

141

Nitric Acid Mfg

30101301 Chemical Manufacturing,
Nitric Acid, Absorber Tail Gas (Pre-1970
Facilities]; 30101302 Chemical
Manufacturing, Nitric Acid, Absorber Tail
Gas (Post-1970 Facilities]

NSCR

$900

6-14 weeks

1,290

724

Petroleum
Refinery
Process Heaters

30600106 Industrial Processes, Petroleum
Industry, Process Heaters, Process Gas-fired

SCR-95%

$940-$l101

28-58 weeks

179

177

12


-------
Gas Turbines

20200201 Natural Gas, Turbine; 20200203
Natural Gas, Turbine: Cogeneration;
20300202 Natural Gas, Turbine

Low NOx Burner
(LNB)

$163

12 months

945

793

Gas Turbines

20200201 Natural Gas, Turbine;
20200203 Natural Gas, Turbine:
Cogeneration; 20300202 Natural Gas,
Turbine; 20300203 Natural Gas, Turbine:
Cogeneration

Low NOx Burner
(LNB]

$800

6-8 months

16,036

4,713

Natural Gas
RICE Pipeline
Compressors

20200202 Internal Combustion Engines,
Industrial, Natural Gas, Reciprocating

Adjust Air to
Fuel Ratio and
Ignition Retard

$249

Uncertain

10,099

2,958

Natural Gas
RICE

Miscellaneous

20100202 Internal Combustion Engines,
Electric Generation, Natural Gas,
Reciprocating; 20200202 Internal
Combustion Engines, Industrial, Natural
Gas, Reciprocating; 20200204, Internal
Combustion Engines, Industrial, Natural
Gas, Reciprocating: Cogeneration;
20300201, Internal Combustion Engines,
Commercial/Institutional, Natural Gas,
Reciprocating

Adjust Air to
Fuel Ratio and
Ignition Retard

$447

Uncertain

27,600

8,085

Natural Gas
RICE Pipeline
Compressors,
Rich Burn

20200253 Internal Combustion Engines,
Industrial, Natural Gas, 4-cycle Rich Burn

NSCR

$517

Uncertain

11,758

10,571

Natural Gas
RICE Pipeline
Compressors,
Lean Burn /
Clean Burn

20200252 Internal Combustion Engines,
Industrial, Natural Gas, 2-cycle Lean
Burn; 20200254 Internal Combustion
Engines, Industrial, Natural Gas, 4-cycle
Lean Burn; 20200255 Internal
Combustion Engines, Industrial, Natural
Gas, 2-cycle Clean Burn; 20200256
Internal Combustion Engines, Industrial,
Natural Gas, 4-cycle Clean Burn

Low Emission

Combustion

(LEC)

$649

Uncertain

47,321

41,169

Diesel / Dual
Fuel RICE

20200401	Internal Combustion Engines,
Industrial, Large Bore Engine, Diesel;

20200402	Internal Combustion Engines,
Industrial, Large Bore Engine, Dual Fuel
(Oil/Gas]

Ignition Retard

$1,255

Uncertain

865

216

13


-------
Catalytic
Cracking (1)

30600201 Industrial Processes, Petroleum
Industry, Catalytic Cracking Units, Fluid
Catalytic Cracking Unit

Low NOx Burner
(LNB) & Flue
Gas Recirculation
(FGR)

$1,375

6-8 months

255

140

Spreader
Strokers

10100204 External Combustion Boilers,
Electric Generation,
Bituminous/Subbituminous Coal,
Spreader Stroker [Bituminous Coal]

SNCR

$1,390

42-51 weeks

394

158

Petroleum
Refinery
Process Heaters

30600106 Industrial Processes, Petroleum
Industry, Process Heaters, Process Gas-fired

SCR-95%

$1,406-$ 1,501

28-58 weeks

161

157

Incinerators

50200102, 50200103, 50200104,
50200504, 30190013, 30190014,
50300101, 50300106, 50300112,
50300113, 50300501, 50300503,
50300504, 50300599, 50100101,

50100102,

50100103,	50100506, 50100515, 50100516,
39990024 Incineration

SNCR

$1,842

42-51 weeks

6,556

2,950

Boilers &
Process Heaters

10200203,10200217, 10300216,10200204,
10200205,10300207,10300209,10200799
External Combustion Boilers; 30190002,
30600103 Industrial Process Heaters

SCR

$2,235

28-58 weeks

13,146

10,358

Natural Gas
RICE Electric
Generation

20100206 Internal Combustion Engines,
Electric Generation, Natural Gas,
Reciprocating: Evaporative Losses (Fuel
Delivery System)

Adjust Air to Fuel
Ratio and Ignition
Retard

$2,347

Uncertain

107

32

Catalytic
Cracking (2)

30600201 Industrial Processes, Petroleum
Industry, Catalytic Cracking Units, Fluid
Catalytic Cracking Unit; 30600202
Industrial Processes, Petroleum Industry,
Catalytic Cracking Units, Catalyst Handling
System

Low NOx Burner
(LNB) & Flue
Gas Recirculation
(FGR)

$2,369

6-8 months

274

97

Gas-Fired
Process Heaters

(1)

30600104 Industrial Processes, Petroleum
Industry, Process Heaters, Gas-fired

SCR-95%

$2,376

28-58 weeks

211

204

Coal Boilers

10200206,10200224,10200225,
10300102,10300208,10300224,
10300225

SNCR

$2,413

42-51 weeks

1099

495

14


-------
Gas-Fired
Process Heaters

30600104 Industrial Processes, Petroleum
Industry, Process Heaters, Gas Fired

Ultra-Low NOx
Burners

$2,419-$2,638

6-8 months

137

64

(2)













By-Product
Coke

Manufacturing

30300306 Industrial Processes, Primary
Metal Production, By-Product Coke
Manufacturing, Oven Underfiring

SNCR

$2,673

42-51 weeks

2,366

1,420

ICI Boilers -
Residual Oil

10200401,10200402, 10200404,10300401,
10300402 External Combustion Boilers,
Residual Oil

LNB & SNCR

$2,850

6-8 months

991

689

Ammonia
Production

30100306 Industrial Processes, Chemical
Manufacturing, Ammonia Production,
Primary Reformer: Natural Gas Fired

SCR

$2,896

28-58 weeks

2,508

2,257

Glass

Manufacturing
-Flat

30501403 Industrial Processes, Mineral
Products, Glass Manufacture, Flat Glass:
Melting Furnace

OXY-Firing

$3,097

Uncertain

9,721

7,880

ICI Boilers

10200201,10200202, 10200212,10300205,
10200501,10200504, 10200601,10200602,
10200603,10200604,10201401,10300601,
10300602,10200701, 10200704,10200707,
10201402 External Combustion Boilers

Low NOx Burner
& SCR

$3,456

6-8 months (LNB)
28-58 weeks (SCR)

31,005

28,204

Iron & Steel -
In-Process
Combustion -

30300819,30300824, 30300913,30300914,
30301522 Industrial Processes, Primary
Metal Production

SCR

$3,705

28-58 weeks

829

746

Bituminous













Coal













Diesel RICE
Miscellaneous

20100102 Internal Combustion Engines,
Electric Generation, Distillate Oil (Diesel],
Reciprocating; 20100107 Internal
Combustion Engines, Electric Generation,
Distillate Oil (Diesel], Reciprocating:
Exhaust; 20200102 Internal Combustion
Engines, Industrial, Distillate Oil (Diesel],
Reciprocating;

20200106 Internal Combustion Engines,
Industrial, Distillate Oil (Diesel],
Reciprocating: Evaporative Losses (Fuel
Storage and Delivery System];

SCR

$3,814

28-58 weeks

1,091

869

15


-------


20200107 Internal Combustion Engines,
Industrial, Distillate Oil (Diesel],
Reciprocating: Exhaust;

20300101 Internal Combustion Engines,
Commercial/Institutional, Distillate Oil
(Diesel], Reciprocating;

20400403 Internal Combustion Engines,
Engine Testing, Reciprocating Engine,
Distillate Oil











Catalytic
Cracking,
Process
Heaters, &
Coke Ovens

30600201, 30390004, 39000701,
39000702,39000797

LNB & FGR

$5,199

6-8 months

1,989

1,094

Petroleum
Refinery Gas-
Fired Process
Heaters (3)

30600104 Industrial Processes, Petroleum
Industry, Process Heaters, Gas-fired,
30600106 Industrial Processes, Petroleum
Industry, Process Heaters, Process Gas-fired

SCR-95%

$8,885-$9,140

28-58 weeks

370

316

Glass

Manufacturing
- Pressed

30501404 Industrial Processes, Mineral
Products, Glass Manufacture, Pressed and
Blown Glass: Melting Furnace

OXY-Firing

$6,356

Uncertain

1,001

851

Petroleum
Refinery Gas-
Fired Process
Heaters (2)

30600104 Industrial Processes,
Petroleum Industry, Process Heaters,
Gas-fired, 30600106 Industrial
Processes, Petroleum Industry, Process
Heaters, Process Gas-fired

SCR-95%

$7,533-$8,120

28-58 weeks

362

304

Industrial
Processes
Miscellaneous

30600201 Industrial Processes,
Petroleum Industry, Catalytic Cracking
Units, Fluid Catalytic Cracking Unit;
39000701 Industrial Processes, In-
process Fuel Use, Process Gas, Coke Oven
or Blast Furnace

LNB & FGR

$4,026

6-8 months

871

479

Glass

Manufacturing
- Container

30501402 Industrial Processes, Mineral
Products, Glass Manufacture, Container
Glass: Melting Furnace

OXY-Firing

$7,481

Uncertain

3,107

2,628

Petroleum
Refinery Gas-

30600104 Industrial Processes,
Petroleum Industry, Process Heaters,
Gas-fired; 30600106 Industrial

SCR-95%

$5,609-$5,884

28-58 weeks

372

338

16


-------
Fired Process
Heaters (1)

Processes, Petroleum Industry, Process
Heaters, Process Gas-fired











Taconite Ore
Processing

30302351, 30302352, 30302359
Industrial Processes, Primary Metal
Production, Taconite Ore Processing,
Induration

SCR

$6,449

28-58 weeks

1,188

991

Diesel RICE

Electric

Generation

20200102 Internal Combustion Engines,
Electric Generation, Distillate Oil (Diesel],
Reciprocating

SCR

$1,499

28-58 weeks

778

622

17


-------
3.4 Discussion ofNon-EGlJ Source Groups

The below discussion utilizes the information in Table 3 in order to assess whether
significant aggregate NOx mitigation is achievable from non-EGU sources by the 2017
ozone season.

It is clear that a number of source categories have been identified by CoST using the least-
cost procedure that have the potential for non-EGU stationary source emissions reductions.
There are some notable source categories that have the potential for substantial non-EGU
stationary source emissions reductions below $10,000 per ton.16 However, for the
purposes of this analysis, the EPA did not further examine control options above $3,400
per ton. This is consistent with the range we analyzed for EGUs in the proposed and final
rules, and is also consistent with what the EPA has identified in previous transport rules as
highly cost-effective, including the NOx SIP call.17 Again, this was done because the
objective of this analysis is to characterize whether significant aggregate NOx mitigation is
achievable from non-EGU sources by the 2017 ozone season, so we focused the search on
categories with highly cost-effective technologies. This focus excludes several source
groups with high emissions reduction potential because reductions from those source
groups are not available for $3,400 per ton or less, including: ICI boilers using SCR & LNB;
Catalytic Cracking, Process Heaters, & Coke Ovens using LNB & FGR; and Pressed and
Container Glass Manufacturing using OXY-Firing.

At a cost level of $3,400 per ton or less, there are a number of source groups with
substantial reduction potential. However the table also identifies several source groups
whose reduction potential is not significant, and which the EPA did not weigh heavily in
assessing the aggregate non-EGU NOx emission reduction potential. This is because the
aggregate potential reductions from these "insignificant" source groups is small. These
"insignificant" source groups comprise those source groups with many small sources, as
well those containing a limited number of larger sources; for either of these types of
groups, potential aggregate emission reductions are small relative to reductions available
from other source categories. The EPA does not believe that small sources have significant
emission reduction potential in the aggregate because most small sources emit less than
100 tons of NOx per year. (It is worth noting that small sources account for a significant
percentage of the total number of non-EGU point sources. See Appendix A/B for more
information on the number of sources within certain states.) The EPA therefore excludes
from the focus of this analysis these insignificant source groups, namely, those with
aggregate potential reductions of 1,000 tons per year or less (which represents less than
0.1 percent of the anthropogenic ozone season inventory).

The EPA will now focus on the several source groups with significant cost-effective
reductions identified in Table 3. These source groups include cement kilns, two types of
cement manufacturing (dry and wet), gas turbines, four separate groups of natural gas
reciprocating IC engines (RICE), incinerators, boilers & process heaters, by-product coke

16	$10,000 per ton represents the cost/ton for Best Available Control Technology (BACT) determinations, which usually
do not exceed $10,000/ton in the Eastern U.S.

17	$3,400 per ton represents the $2,000 per ton value (in 1990 dollars] used in the NOx SIP call, adjusted to the 2011
dollars used throughout this proposal Adjustment of costs was made using the Chemical Engineering Plant Cost Index
(CEPCI] annual values for 1990 and 2011.

18


-------
manufacturing, ammonia production, and flat glass manufacturing. These source groups
are listed below with their control technologies, estimated annualized control costs, and
estimated installation time. These groups have been organized into 7 categories for clarity,
based on either common control technologies (categories 1 through 6) or similarity of
source groups (category 7).

Category 1

-Cement Mfg (dry)

-Incinerators

-By-Product Coke Manufacturing

Category 2

-Cement Kilns

Category 3

-Gas Turbines

Control Tech.

SNCR
SNCR
SNCR

Est. Cost

$1,255
$1,842
$2,673

Biosolid Injection $410
Technology (BSI)

Low NOx Burner $800
(LNB)

Est. Inst. Time

42-51 weeks
42-51 weeks
42-51 weeks

Uncertain

6-8 months

Category 4

-Cement Mfg (wet)

Category 5

-Boilers & Process Heaters
-Ammonia Production

Mid-Kiln Firing $73

SCR
SCR

$2,235
$2,896

5-7 months

28-58 weeks
28-58 weeks

Category 6

-Glass Manufacturing - Flat
Category 7

-Gas RICE Pipeline Compressors

-Gas RICE Miscellaneous

-Gas RICE Pipeline Compressors,
Rich Burn

-Gas RICE Pipeline Compressors,
Lean/Clean Burn

OXY-Firing

$3,097

Adjust AFR and $249
Ignition Retard
Adjust AFR and $447
Ignition Retard
NSCR	$517

Low Emission $649
Combustion (LEC)

Uncertain

Uncertain
Uncertain
Uncertain
Uncertain

The EPA makes the following observations about the potential reductions from these
significant cost-effective categories.

The source groups listed in Category 1 would utilize SNCR as the recommended control
technology. The time necessary to install SNCR equipment is generally well known. A
typical installation timeline of 42-51 weeks is generally needed to complete a SNCR project

19


-------
going from the bid evaluation through startup, which installation timeline is specific to
non-EGUs. Based on this fact alone (which does not consider additional time likely
necessary for permitting or installation of monitoring equipment), the ability for SNCR
technology to be installed and operational in time for the 2017 ozone season seems very
unlikely.

The source group listed in Category 2 contains a specific source of uncertainty in regards to
biosolid injection technology (BSI). Due in large part to the lack of widespread use of this
control technology, research performed by the EPA has been unable to uncover any reliable
information on the time required to install the necessary BSI equipment on cement kilns.
Compliance timing with regard to biosolid injection technology should therefore be
considered extremely uncertain. Based on this fact alone (and aside from additional time
likely necessary for permitting or installation of monitoring equipment), the ability for this
technology to be installed and operational at all facilities in this category in time for the
2017 ozone season is unknown.

The source group listed in Category 3 would utilize LNB as the recommended control
technology, with a necessary installation time of approximately 6-8 months. Some of the
LNB combustion control technology identified for non-EGU sources reflects a different
technology that may have different timing considerations than that considered for EGU
boilers. For instance, LNB at non-EGU combustion turbines in this assessment refers to
"dry low-NOx burners" (DLNB) which, in addition to the usual diffusion burner, typically
also include provisions to "premix" natural gas and combustion air prior to combustion. In
spite of the similarity in naming, this is a different technology than the LNB technology
examined and assumed for reductions at EGU boilers. Therefore, the same timing
assumptions assumed and demonstrated on the EGU side are not necessarily applicable to
combustion control technology for non-EGU sources. Moreover, non-EGUs of any type -
boiler or turbine - that are not currently required to monitor and report in accordance
with 40 CFR Part 75 will require additional time relative to EGUs that are currently
equipped with Part 75 monitoring and reporting (such as those EGUs covered under
federal transport rulemakings and this one). Installation of NOx monitors for the reporting
of NOx mass requires the construction of platforms, Continuous Emissions Monitoring
(CEM) shelters, procurement of equipment, certification testing, and electronic data
reporting programming of a data handling system. These timing considerations on the
non-EGU sources make installation of controls by the 2017 timeframe established in this
rule less likely and more uncertain for industrial sources.

The source group listed in Category 4 would utilize mid-kiln firing as the recommended
control technology. A fairly well-known aspect is the time necessary to install this
equipment; typically, 5-7 months is needed to complete a mid-kiln firing project going from
the bid evaluation through startup. However, the above-discussed issues regarding
monitoring and reporting of NOx mass on non-EGU sources that currently lack such
monitoring equipment make installation of controls by the 2017 timeframe of this rule less
likely and more uncertain for industrial sources such as those in the cement manufacturing
(wet) source group.

20


-------
The source groups listed in Category 5 would utilize SCR as the recommended control
technology, with an installation time of 28-58 weeks for SCR (dependent on exhaust gas
flow rates; larger systems require longer installation times). Based on the installation time
frame alone (which does not consider additional time likely necessary for permitting or
installation of monitoring equipment), the ability for SCR technology to be installed and
operational in time for the 2017 ozone season seems unlikely In addition to this
uncertainty, the above-discussed issues regarding monitoring and reporting of NOx mass
on non-EGU sources that currently lack such monitoring equipment make installation of
controls by the 2017 timeframe established in this rule less likely and more uncertain for
industrial sources such as those in Category 5 source groups.

The source group listed in Category 6 would utilize OXY-Firing as the recommended
control technology, with an uncertain necessary installation. A specific source of
uncertainty with regard to the estimated installation time of this control technology is that
OXY-Firing is generally installed only at the time of a furnace rebuild, which rebuilds may
occur at infrequent intervals of a decade or more.18 In addition to this uncertainty, the
above-discussed issues regarding monitoring and reporting of NOx mass on non-EGU
sources that currently lack such monitoring equipment make installation of controls by the
2017 timeframe established in this rule less likely and more uncertain for industrial
sources such as those in Category 6 source group.

Finally, the source groups listed in Category 7 are all RICE. While some of the
recommended control technologies may involve installation timelines that are relatively
short on a per-engine basis, there is substantial uncertainty in large-scale installation over
numerous sources. References indicate that implementation of NOx controls of any type on
a large number of RICE will require significant lead time to train and develop resources to
implement emission reduction projects; market demand could significantly exceed the
available resource base of skilled professionals.19 Additionally, in order not to disrupt
pipeline capacity, engine outages must be staggered and scheduled during periods of low
system demands for those engines involved in natural gas pipelines (as is the case with 3 of
the 4 RICE source groups with significant cost-effective reductions). In addition to this
uncertainty, the above-discussed issues regarding monitoring and reporting of NOx mass
on non-EGU sources that currently lack such monitoring equipment make installation of
controls by the 2017 timeframe established in this rule less likely and more uncertain for
industrial sources such as RICE.

4 Summary of Comments Receive '	ted Rule TSD

The EPA received relatively few comments on the draft Assessment of Non-EGU NOx
Emission Controls, Cost of Controls, and Time for Compliance TSD provided in the docket
for this rule. None of these comments changed our conclusions reached in the draft TSD, as
commenters generally agreed with the EPA's assessment with respect to the regulation of

18	See Appendix B.

19	"Availability and Limitations of NOx Emission Control Resources for Natural Gas-Fired Reciprocating Engine Prime
Movers Used in the Interstate Natural Gas Transmission Industry," Innovative Environmental Solutions Inc., July 2014.

21


-------
non-EGUs in this rule. Detailed responses to these comments can be found in the response
to comments document available in the docket for this final rule. A brief discussion of one
comment containing data on control information is presented below.

Commenter Fuel Tech, Inc. (FTI) provided information on the installation time required for
SNCR equipment, stating that "FTI has provided SNCR systems in 8-12 months (from
contract award to performance guarantee certification) ..."20 This timeframe is largely
consistent with the 42-51 week (9.7-11.7 month) timeframe estimate presented by the EPA
in the draft TSD.

In addition, FTI provided information on a range of SNCR cost per ton based on
installations from 2010 to 2015 on non-EGU sources, stating that "these recent examples
show NOx reduction cost effectiveness in the range of $2,200 to $2,900 per ton of NOx
removed on an annual basis."21 FTI's "Figure l"22 also provided a chart of cost effectiveness
($/ton) versus unit size (mmBTU/hr) for both annual and ozone season NOx. A log fit of the
ozone season curve shows cost effectiveness in the range of approximately $2,000 to
approximately $6,500 per ton of ozone season NOx removed, with installations tending to
be more expensive for smaller unit sizes. Although FTI's estimates are based on different
interest rates and capital investments than our estimates, they are worthwhile to note in
comparison to our stated estimate of $1,300 to $2,700 per ton of NOx removed on an ozone
season basis.

5 Conclusion

The above preliminary analysis performed by the EPA indicates that uncertainty exists
regarding whether significant aggregate NOx mitigation is achievable from non-EGU point
sources by the 2017 ozone season. Reducing this uncertainty requires further
understanding of potentially available control measures that could have annualized costs of
$3,400 per ton or less. In addition, further implementation of the recommendations in the
Appendices to this TSD, the extent of which as determined by the EPA to be needed, may
also reduce our uncertainty regarding the credibility of data for control measures included
in future non-EGU NOx control strategy efforts. Please note that while the information in
these Appendices supports our conclusion regarding whether significant aggregate NOx
mitigation is achievable from non-EGU point sources by the 2017 ozone season, this final
TSD is making no conclusions about the recommendations for further improvements.

While a number of source groups with control options were identified, the EPA did not
further examine control options above $3,400 per ton, consistent with the range analyzed
for EGUs in the proposed and final rules and with what the EPA has identified in previous
transport rules as highly cost-effective. A number of source groups were identified at a cost
level of $3,400 per ton or less, however the EPA believes several of these source groups
may not be significant. Of the remaining source groups, a variety of factors indicated the
ability for control technology to be installed and operational in time for the 2017 ozone

20	EPA-HQ-OAR-2015-0500-0356, page 3.

21	EPA-HQ-OAR-2015-0500-0356, page 7.

22	EPA-HQ-OAR-2015-0500-0356, page 6.

22


-------
season seemed unlikely, with an overarching consideration being that non-EGUs of any
type that are not currently required to monitor and report in accordance with 40 CFR Part
75 will require additional time for implementation relative to EGUs that are currently
equipped with Part 75 monitoring and reporting. These added timing considerations on the
non-EGU sources make installation of controls by the 2017 timeframe established in this
rule less likely and more uncertain for industrial sources.

With all of these factors being considered, the limited available information points to an
apparent scarcity of non-EGU reductions that could be accomplished by the beginning of
the 2017 ozone season. As noted in the proposed and final rule, this conclusion has led the
EPA to focus the final FIPs on EGU reductions. Both the proposal and the final rule
acknowledge that this may not be the full remedy that is ultimately needed to eliminate an
upwind state's significant contribution to nonattainment or interference with maintenance
of the 2008 ozone NAAQS (or, for that matter, the 2015 ozone NAAQS) in other states.
Emissions reductions from the non-EGU categories discussed above may be necessary,
though on a longer timeframe than the 2017 compliance deadline being finalized in this
rulemaking. The EPA intends to explore this question further in future ozone transport
rulemakings.

23


-------
May 2014

Update of NOx Control Measure Data in the
CoST Control Measure Database for Four

Industrial Source Categories:
Ammonia Reformers, NonEGU Combustion
Turbines, Glass Manufacturing, and Lean Burn
Reciprocating Internal Combustion Engines

Final Report

Prepared for

Larry Sorrels

U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards

Air Economics Group
Research Triangle Park, NC 27711

Prepared by

RTI International

3040 E. Cornwallis Road
Research Triangle Park, NC 27709

RTI Project Number 0212979.002.002

HRTI

INTERNATIONAL


-------
RTI Project Number
0212979.002.002

Update of NOx Control Measure Data in the
CoST Control Measure Database for Four

Industrial Source Categories:
Ammonia Reformers, NonEGU Combustion
Turbines, Glass Manufacturing, and Lean Burn
Reciprocating Internal Combustion Engines

Revised Draft Report

October 2014

Prepared for

Larry Sorrels

U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards

Air Economics Group
Research Triangle Park, NC 27711

Prepared by

RTI International

3040 E. Cornwallis Road
Research Triangle Park, NC 27709

RTI International is a trade name of Research Triangle Institute.


-------
CONTENTS

Section	Page

1	Introduction	1-1

2	Ammonia Reformers Sector	2-1

2.1	Recommended Deletions	2-2

2.2	Recommended Additions	2-2

2.3	Recommended Changes	2-3

2.4	Other Comments	2-5

3	Combustion Turbines	3-1

3.1	Recommended Del eti on s	3-2

3.2	Recommended Additions	3-2

3.2.1	Catalytic Combustion; Gas Turbines—Natural Gas
(NCATCGTNG)	3-3

3.2.2	EMx and Water Injection; Gas Turbines—Natural Gas
(NEMXWGTNG)	3-4

3.2.3	EMx and Dry Low NOx Combustion; Gas Turbines—Natural Gas
(NEMXDGTNG)	3-8

3.3	Recommended Changes	3-9

3 3 I Water Injection; Gas Turbines—Natural Gas (NWTINGTNG)	3-10

3.3.2	Steam Injection; Gas Turbines—Natural Gas (NSTINGTNG)	3-11

3.3.3	Dry Low NOx Combustion; Gas Turbines—Natural Gas
(NDLNCGTNG)	3-12

3.3.4	SCR and Water Injection; Gas Turbines—Natural Gas
(NSCRWGTNG)	3-13

3.3.5	SCR and Steam Injection; Gas Turbines—Natural Gas
(NSCTSGTNG)	3-16

3.3.6	SCR and Dry Low NOx Combustion; Gas Turbines—Natural Gas
(NSCRDGTNG)	3-17

3.3.7	Water Injection; Gas Turbines—Oil (NWTINGTOL)	3-19

in


-------
3.3.8	SCR and Water Injection; Gas Turbines—Oil (NSCRWGTNG)	3-20

3.3.9	Water Injection; Gas Turbines—Jet Fuel (NWTINGTJF)	3-22

3.3.10	SCR and Water Injection; Gas Turbines—Jet Fuel (NSCTWGTJF).... 3-22

3.3.11	Applicable Control Measures for Gas Turbine SCCs	3-22

3.4	Example Emission Limits for NonEGU Combustion Turbines	3-26

3.5	References			3-26

4	Glass Manufacturing Sector	4-1

4.1	Introduction			4-1

4.2	Example NOx Regulatory Limits	4-1

4.2.1	Wisconsin	4-1

4.2.2	New Jersey	4-1

4.2.3	New York	4-1

4.3	Recommended Additions	4-1

4.4	Recommended Changes	4-3

4.5	Recommended Deletions	4-5

4.6	Updates to Source Classification Codes	4-5

4.7	References		4-7

5	Lean Burn Engines..		5-1

5.1	Literature Search	5-1

5.2	Document Review	5-2

5.3	Low Emission Combustion (LEC) (NLECICENG)	5-3

5.4	Layered Combustion (LC), 2 Stroke (NLCICE2SNG)	5-5

5.5	Layered Combustion (LC), Large Bore, 2 Stroke, Low Speed
(NLCICE2SLBNG)	5-7

5.6	Air to Fuel Ratio Controller (AFRC) (NAFRCICENG)	5-9

5.7	SCR (for 4 Stroke Natural Gas Engines) (NSCRICE4SNG)	5-10

iv


-------
5.8	SCR (for Diesel Engines) (NSCRICEDS)	5-12

5.9	Applicable SCCs for Lean Burn Engine Control Measures	5-14

Appendixes

A Ammonia Reformers	A-l

B Combustion Turbines	B-l

C Glass Manufacturing		.C-l

D Lean Burn Engines			D-l

E Notes Provided Here to EPA Questions on Lean Burn RICE	E-l

v


-------
LIST OF TABLES

Number	Page

2-1 Applicable SCCs for the Ammonia Production Industry	2-4

2-2	RACT NOX Limits for ICI Natural Gas Boilers	2-5

3-1	Summary of Cost Effectiveness and Supporting Data for Catalytic Combustion	3-4

3-2 Summary of Cost Effectiveness and Supporting Data for EMx Plus Water

Injection	3-7

3-3 Summary of Cost Effectiveness and Supporting Data for EMx Plus Dry Low

NOx Combustion	3-8

3-4 Summary of Cost Effectiveness and Supporting Data for DLN Combustion	3-13

3-5 Summary of Cost Effectiveness and Supporting Data for SCR Plus Water

Injection	3-15

3-6 Summary of Cost Effectiveness and Supporting Data for SCR Plus Steam

Injection (SI)	3-17

3-7 Summary of Cost Effectiveness and Supporting Data for SCR Plus DLN

Combustion	3-18

3-8 Summary of Cost Effectiveness and Supporting Data for SCR Plus Water

Injection (WI) for Oil-Fired Turbines	3-21

3-9 Recommended Control Measures for Gas Turbine SCCs	3-23

3-10	NOx Emissions Limits for NonEGU Combustion Turbines in New York	3-26

4-1	Summary of Cost Effectiveness and Supporting Data for Recommended

Additions	4-2

4-2 Summary of Cost Effectiveness and Supporting Data for Recommended

Additions	4-5

4-3	Applicable SCCs for the Glass Manufacturing Industry	4-6

5-1	LEC for Natural Gas Lean Burn Engines	5-4

5-2 LC for Natural Gas Lean Burn Engines, 2-stroke	5-6

5-3	LC for Natural Gas Lean Burn Engines, Large Bore 2-stroke	5-8

5-4 AFRC for Natural Gas Lean Burn Engines	5-9

5-5	SCR for Natural Gas Lean Burn Engines, 4-stroke	5-11

5-6 SCR for Diesel Lean Burn Engines—Assumptions	5-13

5-7 Potential Reciprocating Engine SCCs to Add to the CMDB and Applicable

Control Measures	5-15

5-8 Recommended New Control Measures to Associate With Lean Burn

Reciprocating Engine SCCs in the CMDB	5-17

5-9 Recommended Control Measure Deletions From SCCs in the CMDB	5-18

5-10 NOx Control Requirements for RICE in Pennsylvania	5-21

5-11 Characteristics of NOx Emissions and Controls for RICE	5-21

vi


-------
SECTION 1
INTRODUCTION

The U.S. Environmental Protection Agency (EPA) Health and Environmental Impacts
Division (HEID) has developed the Control Strategy Tool (CoST) to support national- and
regional-scale multipollutant air quality modeling analyses. CoST allows users to estimate the
emissions reductions and costs associated with future-year emission control strategies, and then
to generate emission inventories that reflect the effects of applying the control strategies. The
tool uses EPA HEID's Control Measures Database (CMDB) to develop control strategies and
provides a user interface to that database. The CMDB is a relational database that contains
information on an extensive set of control measures for point sources, nonpoint sources, and
mobile sources. Information contained in the database includes descriptions of the measures,
control efficiencies for the pollutants affected, costs of control, and the types of sources or
processes to which the control measures can be applied. The database includes robust cost
equations to determine engineering costs for some control measures that take into account how
control costs vary with respect to variables for the source such as unit size or flow rate. The
database also includes simple cost factors for all source types in terms of dollars per ton of
pollutant reduced that can be used to calculate the cost of the control measure if the applicable
source variable data are unavailable or no equation has been developed.

This report presents the results of an effort to review and enhance the CMDB with new and/or
updated NOx control measure data for the following four industrial source categories: ammonia
reformers, combustion turbines (nonEGU), glass manufacturing, and lean burn reciprocating
internal combustion engines. Section 2 of this report describes the procedures used to locate
more recent data than that currently in the CMDB for control measures applicable to ammonia
reformers. Section 2 also identifies the source of the new data, describes any modifications to the
assumptions or procedures in the referenced analyses needed to make the results consistent with
results for other control measures in the database (such as operating hours for determination of
total annual costs), and describes the specific recommended changes or additions to the database.
Sections 3 through 5 of this report provide similar details for combustion turbines, glass
manufacturing, and lean burn reciprocating internal combustion engines, respectively. Appendix
A presents all of the records for ammonia reformer control measures in each of the CMDB tables
showing their content after making the recommended revisions described in the report.
Appendixes B through D provide comparable tables for the combustion turbine, glass
manufacturing, and lean burn reciprocating internal combustion engine (RICE) source
categories, respectively. Appendix E provides answers to questions on lean-burn RICE NOx

1-1


-------
emissions and available control measures. It should be noted that these revisions and updates
will improve the accuracy and quality of NOx non-EGU control strategy and cost analyses for
EPA rulemakings.

1-2


-------
SECTION 2
AMMONIA REFORMERS SECTOR

The control measures database includes the following NOx emissions control measures
for ammonia reformers:

¦	Oxygen trim and water injection,

¦	Low NOx burners and flue gas recirculation,

¦	Selective non-catalytic reduction (SNCR),

¦	Selective catalytic reduction (SCR), and

¦	Low NOx burners.

In order to update the existing control measures database. a lilcialuic search was
conducted using the following terms:

¦	reformer

¦	cost

¦	"NOx" or "nitrogen oxide"

¦	"Low NOx burner" or "LNB"

¦	"Flue gas recirculation" or "FGR"

¦	oxygen trim

¦	water injection

¦	"Selecth e catalytic reduction" or "SCR"

¦	"Selective non catalytic reduction" or "SNCR"

¦	emission reduction

¦	control efficiency

Due to the use of SCR and SNCR to control NOx emissions and the fact that ammonia is
used in the operation of SCR and SNCR, the literature search resulted in NOx reductions on
processes other than ammonia production.

2-1


-------
In order to focus on ammonia production, a focused internet search for operating permits,
BACT analyses, and NOx controls was conducted using the 22 ammonia production facilities in
the United States.

As a result of the following facts, most of the internet search results included NOx
reductions from the production of nitric acid production instead of ammonia production:

¦	The NOx emissions from nitric acid production are covered by a New Source
Performance Standard (NSPS) codified as Subpart G and Subpart Ga of Pan 60

¦	Nitric acid facilities covered by the NSPS are required to install NOx continuous
emission monitoring systems (CEMS).

¦	Many nitric acid facilities use SCR to control NOx emissions.

¦	Many ammonia production facilities are co-located with nitric acid production
facilities.

The internet search resulted in one new NOx reduction project, which was the result of a
voluntary agreement between Terra Nitrogen and the Indian Nations Council of Governments to
install "ultra-low NOx burner technology to an existing ammonia reformer [and] reduce the
unit's NOx emissions by approximately 60% at a projected capital cost of two million dollars."
The existing ammonia reformer is located at Terra Nitrogen, L.P., Verdigris Plant in Claremore,
Oklahoma.

Based on information known to EPA and collected for this report, Low NOx burner
technologies are known and demonstrated control techniques for ammonia reformers.

The following sections outline the deletions, additions, changes, and other comments
recommended for the CMDB in relation to NOx emissions from ammonia reformers.

2.1	Recommended Deletions

No deletions are recommended.

2.2	Recommended Additions

The only addition to the CMDB is to add the following reference: Tulsa Metropolitan
Area 8-Hour Ozone Flex Plan: 2008 8-O3 Flex Program. Prepared by Indian Nations Council of
Governments (INCOG), 201 W. 5th Street, Suite 600, Tulsa, OK 74103. March 6, 2008.
http://www.epa.gov/ozoneadvance/pdfs/Flex-Tulsa.pdf.

This addition is shown in Appendix A as Table A-l.

2-2


-------
2.3 Recommended Changes
Updates to costs and Efficiencies.

Changes to one record (LNB applied to large source types) are recommended to reflect
the new reference dated March 6, 2008.

Using the new reference and a reference already contained in the CMDIi. I he following
assumptions were made:

¦	NOx reductions of 425 tons per year1

¦	Capital cost of $2 million1

¦	Maintenance costs are 2.75% of capital costs

¦	Equipment life of 10 years

¦	Interest rate of 7%

¦	Capital recovery factor of 0.1424

The resulting annual costs are $339,800 and the cost effectiveness is $800 per ton of NOx
reduction (both in 2008 dollars). The capital to annual cost ratio is 5.9.

The previous entry showed a cost effectiveness of $650 per ton of NOx reduction (in
1990 dollars) and a capital to annual cost ratio is 5.5. The changes are included in Appendix A as
Table A-2 and Table A-3. Changes are indicated by red, italic text.

Updates to Source Classification Codes.

The U.S. Environmental Protection Agency (USEPA) developed the Source
Classification Code (SCC) system, which assigns an eight digit code to each emission unit based
on the general criteria pollutant emission point type, the major industry group, specific industry
group, and specific process unit/fuel combination. The system allows similar emission points to
be grouped together for analyses.

For ammonia reformers, there are seven applicable SCCs, as shown in Table 2-1.

1	Indian Nations Council of Governments (INCOG), 2008: Indian Nations Council of Governments (INCOG),
"Tulsa Metropolitan Area 8-Hour Ozone Flex Plan: 2008 8-03 Flex Program," March 6, 2008. Downloaded from
http://www.epa.gov/ozoneadvance/pdfs/Flex-Tulsa.pdf.

2	U.S. Environmental Protection Agency. Alternative Control Techniques Document— NOx Emissions from
Process Heaters (Revised), document EPA-453/R-93-034, dated September 1993.

2-3


-------
Table 2-1. Applicable SCCs for the Ammonia Production Industry

SCC

SCC1

SCC3

SCC6

SCC8

30100305

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Feedstock
Desulfurization

30100306

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Primary Reformer:
Natural Gas Fired

30100307

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Primary Reformer: Oil
Fired

30100308

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Carbon Dio\nle
Regenerator

30100309

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Condensate Stripper

30100310

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Storage and Loading
Tanks

30100399

Industrial Processes

Chemical
Manufacturing

Ammonia Production

Other Not Classified

In an analysis of NOx emissions for the Ozone Transport Region in 2011, four of the
SCCs in Table 2-1 were identified. These SCCs are 30100306, 30100307, 30100310, and
30100399. Only SCCs 30100306 and 30100307 are associated with ammonia reformer NOx
controls in the current CMDB.

The known control techniques for ammonia reformers are typically used for point
emission sources, such as stacks. Emissions from SCC 30100310 are not typically vented, so
capture and control of these emissions is likely not feasible. Therefore, no changes related to
SCC 301003 10 are recommended lor the CMDB.

For 1 Ik- purposes of this analysis, SCC 30100399 is assumed to include combustion
emissions from gaseous fuels other than natural gas. Therefore, all control techniques that are
applicable to natural gas fired ammonia reformers are assumed to also apply to SCC 30100399.
Also, the cost to control NOx emissions from gaseous fuels is assumed to be comparable to the
cost to control NOx emissions from natural gas. Therefore, the costs related to those control
techniques are assumed to apply to SCC 30100399.

The applicable SCC from Table 2-1 was added to the Description field for each control
technique in Table A-2 of Appendix A. SCC 30100306 was already included in the table; SCCs
30100305, 30100307, and 30100399 were added, where appropriate. Changes are indicated by
red text.

2-4


-------
2.4 Other Comments

Control measures used by ICI boilers. Review of NOx control measures used for
boilers was not included in this analysis. However, many of the SCR costs in the CMDB for
natural gas fired ammonia reformers are based on SCR costs for

Industrial/Commercial/Institutional (ICI) Boilers using Process Gas. No SCR costs specific to
ammonia reformers were noted in the CMDB.

At a later time, it may be pertinent to review recent final ICI Boilers ivuulalions or other
sources for potential updates to the cost of SCR on ammonia reformers. The final major source
NESHAP for ICI Boilers was promulgated on January 31, 2013 and the final area source
NESHAP for ICI Boilers was promulgated on February 1, 2013.

Potential NOx limits for ammonia reformers based on boiler NOx limits. According
to NOx Reasonably Acceptable Control Technology (R.ACT), the states of New Jersey and New
York have established emission limits for ICI Natural Gas Boilers (greater than 100 million BTU
per hour) that could be applicable to natural gas ammonia reformers. These R.ACT limits are
shown in Table 2-2.

Table 2-2. RACT NOx Limits for ICI Natural Gas Boilers

State

Boiler Size

Limit (lb NOx/MMBTU)

Effective Date

New Jersey3

>100 MMBTl

0.10

Already in effect

New York

>100 MMBTl and :5o \1\1l 5TIJ

0.06

7/1/14

New York

>250 MM 1 i l l

0.08

7/1/14

a The limil also applies in oilier iih.Iiiv.vi heal exchangers.

2-5


-------
SECTION 3
COMBUSTION TURBINES

The CMDB includes the following NOx emissions control measures for Combustion
Turbines:

¦	Water injection for natural gas-fired turbines (achieves 76 percent reduction)

¦	Steam injection for natural gas-fired turbines (achieves 80 percent reduction)

¦	Low NOx Burners for natural gas-fired turbines (achieves 84 percent reduction)

¦	SCR on natural gas-fired turbines that also have water injection (achieves 95 percent
reduction)

¦	SCR on natural gas-fired turbines that also have steam injection (achieves 95 percent
control)

¦	SCR on natural gas-fired turbines that also have low NOx burners (achieves 94
percent reduction)

¦	Water injection for oil-fired turbines (achieves 68 percent reduction)

¦	SCR on oil-fired turbines that also ha\ e ualer i njection (achieves 90 percent
reduction)

¦	Water injection for jet fuel-fired turbines (achieves 68 percent reduction)

¦	SCR on jet fuel-fired turbines that also have water injection (achieves 90 percent
reduction)

¦	Water injection lor aeroderivative turbines (achieves 40 percent reduction)

All of the cost data are in 1990 dollars, except the costs of water injection for
aeroderiv ali\ e turbines, which are in 2005 dollars. In addition, all of the costs are based on
estimated operation for 8,000 hr/yr, except the costs of water injection for aeroderivative
turbines, which are for intermittently operated units. The costs in 1990 dollars are based
primarily on analyses in EPA's 1993 ACT document for NOx Emissions from Stationary Gas
Turbines (EPA, 1993). Capital and annual cost equations are provided for all of the controls
except those for jet fuel-fired turbines and water injection for aeroderivative turbines.

Literature search. In order to update the existing CMDB, a literature search was
conducted for articles and papers published since 2008. In addition, an internet search was
conducted for BACT analysis reports and control technology reports prepared for federal and

3-1


-------
state agencies and RPOs. The literature search did not identify any documents with cost data, but
the internet search identified the documents listed in Section 3.5 of this report.

Changes to CMDB. The following sections outline the deletions, additions, and other
changes recommended for the CMDB in relation to NOx emissions from Combustion Turbines.
All cost data and calculations are in an Excel Worksheet (RTI, 2014). Copies of the CMDB
tables with recommended revisions to the records for combustion turbine controls are provided
in Appendix B.

The coefficient of determination (R2) is 1.0 for many of the regression equations
presented in sections 3.2 and 3.3. The R2 value is exactly 1.0 in cases where the analysis was
based on only two data points; these cases are noted in the discussions for the particular control
measure. In other cases, actual R2 values greater than 0.995 have been rounded to 1.0. These
high values likely are due to the fact that available data for most control measures are from a
single source, and those sources may have already developed a correlation and then picked
specific data points from that correlation for presentation in their documentation.

3.1	Recommended Deletions

RTI recommends deleting the record lor w tiler injection for aeroderivative turbines
because the estimated costs are for combustion turbines that operate on a limited and intermittent
basis (i.e., peaking EGUs). In principle, data for small EGU combustion turbines would be
acceptable for estimating costs of control measures for nonEGUs. However, the limited
operation of peaking units is inconsistent with the assumed operating time of about 8,000 hr/yr
for all of the other nonEGU combustion turbine control measures in the database. For several
SCCs that are currently associated with this control measure in the CMDB we are recommending
applying other existing control measures, as discussed in Section 3.3.11 of this report.

The CMDB also currently applies several gas turbine control measures to reciprocating
internal combustion engine SCCs and to gas turbine SCCs for evaporative losses from fuel
storage and delivery systems. We recommend deleting these applications of the gas turbine
control measures, as discussed in Section 3.3.11 and Section 5.9 of this report.

3.2	Recommended Additions

There are 3 control technique additions for emerging technologies to be added to the
CMDB; these additions include:

¦ Catalytic Combustion; Gas Turbines—Natural Gas;

3-2


-------
¦	EMx and Water Injection; Gas Turbines—Natural Gas;

¦	EMx and Dry Low NOx Combustion; Gas Turbines—Natural Gas.

3.2.1 Catalytic Combustion; Gas Turbines—Natural Gas (NCATCGTNG)

Catalytic combustion is a flameless process that allows fuel oxidation to occur at
temperatures approximately 1800°F lower than those of conventional combustors (OSEC, 1999).
Lower temperatures are desirable because NOx emissions levels are strongly correlated with
temperature. One design that has been commercialized is the Xonon™ combustor (now called K-
Lean™). In the Xonon combustor, a small amount of fuel is burned in a low temperature pre-
combustor. Additional fuel is then mixed with the air and combustion gases from the pre-
combustor and passed through a catalyst module. The catalyst promotes a flameless reaction
between some of the fuel and oxygen. The gases then enter a burnout zone in which the
remaining fuel burns. The maximum temperature in the system is between 2300°F and 2700°F.
In addition to low NOx emissions, the catalytic combustor generates very little CO emissions.
(Peltier, 2003; CARB, 2004; Leposky, 2004; Kawasaki, 2010; Quackenbush, 2012)

Since 1999 at least six Xonon combustors ha\ e been installed; all are 1.4 MW units
(CARB, 2004; Kawasaki, 2010; Quackenbush, 2012). Testing of four of the operating Xonon
combustors has shown NOx emissions less than 3 parts per million by volume on a dry basis
(ppmvd) at 15% oxygen, and permit limits range from 3 ppmvd to 20 ppmvd at 15% oxygen
(CARB, 2004; Quackenbush, 2012). Several companies have conducted research into developing
larger catalytic combustors and other types of designs, but no information was found indicating
that such units have been commercialized (CARB, 2004; Leposky, 2004; Cybulski, 2006).

Although one type of catalytic combustor has been commercialized, we recommend
considering catalytic combustion as an emerging technology in the CMDB because so few units
are in operation, and they are all only one size. In addition, as of 1999, issues with catalytic
combustors include the need for the air-fuel mixture to have completely uniform temperature,
composition, and velocity profile to assure effective use of all the catalyst and to prevent damage
to the substrate from high temperatures. Also the catalyst durability is uncertain (OSEC, 1999).

The recommended costs are based on costs presented in a report by Onsite Sycom Energy
Corporation (OSEC, 1999). The only change we made to the OSEC costs was to calculate capital
recovery using an interest rate of 7 percent instead of 10 percent; this change makes the capital
recovery costs consistent with guidance in Circular A-4 from the Office of Management and
Budget. Table 3-1 summarizes the recommended cost effectiveness and capital to annual cost
ratios for implementing the catalytic combustion NOx control technology. With an outlet

3-3


-------
concentration of 3 ppmvd, catalytic combustion achieves an average reduction of 98 percent
relative to uncontrolled conventional diffusion combustion.

Table 3-1. Summary of Cost Effectiveness and Supporting Data for Catalytic Combustion

Turbine
Output, MW

Cost
Year

Uncontrolled NOx
Emissions

Avg. ppmvd tpy

Outlet
- Concentration,
ppmvd

Cost
Effectiveness,
$/ton NOx

Capital to
Annual Cost
Ratio

Small (5.2)

1999

150

<365

3

920

1.7

Small (26.3)

1999

130

>365

3

670

1.2

Large (170)

1999

210

>365

3

370

0.7

Based on regression of the data in the analysis, the best fit trend lines are represented by
the following equations for the uncontrolled scenario:

Total capital investment (1999 dollars) = 20668 x (MMBtu/hr) A ° '7 (R2=1.0)

Total annual cost (1999 dollars) = 4254.2 x (MMBtu/hr) A 0 82 (R2=1.0)

For all but the smallest turbines, the incremental cost of catalytic combustors relative to
conventional combustors is less than the incremental cost of DLN combustion versus
conventional combustors. Thus, there are no incremental capital costs for catalytic combustion
relative to conventional combustion. However, there are incremental annual costs because the
cost of catalyst replacement is high A best fit equation for incremental catalytic combustion total
annual costs relative to a RACT baseline of DLN combustion is:

Total annual cost (1999 dollars) = 743.22 x (MMBtu/hr) + 54105 (R2=1.0)
3.2.2 EMx and Water Injection; Gas Turbines—Natural Gas (NEMXWGTNG)

Like SCR, EMx™ (formerly called SCONOX™) is a post-combustion catalytic NOx
reduction technology. EMx uses a precious metal catalyst and a NOx absorption/regeneration
process to convert CO and NOx to C02, H20, and N2. NOx reacts with the potassium carbonate
absorbent coating the surface of the oxidation catalyst in the EMx reactor, forming potassium
nitrites and nitrates that are deposited onto the catalyst surface. Each segment, or "can," within
the reactor becomes saturated with potassium nitrites and nitrates over time and must be
desorbed. Regeneration is accomplished by isolating the can via stainless steel lovers and

3-4


-------
injecting hydrogen diluted with steam. Hydrogen is generated onsite with a small reformer that
uses natural gas and steam as input streams. The hydrogen concentration of the reformed gas is
typically 5 percent. Hydrogen and carbon dioxide react with the potassium nitrites and nitrates to
form N2 and H20 and to regenerate the potassium carbonate for another absorption cycle.
(OSEC, 1999; CARB, 2004)

At least 8 EMx systems at 6 facilities have been installed on combustion turbines with
capacities up to 45 MW. Permit limits at most of these facilities have been set al 2 5 ppm vd for
gas-fired operation. EPA has certified it as "demonstrated in practice" LAER-le\ el technology
that reduces NOx to less than 5 ppmvd. The operating range of the catalyst is 300 to 700°F,
which means the technology is not applicable for simple cycle turbines. The vendor for the
technology has indicated that these systems also reduce carbon monoxide emissions to
undetectable levels (essentially 100 percent reduction), reduce volatile organic compound
emissions by greater than 90 percent, and reduce fine particulate matter emissions by 30 percent
(EmeraChem, 2004). Test data documenting these reductions are not available. For the purposes
of the CMDB database, we recommend that this control measure be listed as an emerging
technology (rather than known) because its use has been limited to only a few small turbines.

The recommended costs for EMx in the combined EMx/water injection control measure
are based on costs presented in a 2008 cost estimate prepared by EmeraChem Power for the Bay
Area Air Quality Management District (ECP, 2008). For the purposes of developing 2008 cost
inputs for the CMDB, we made the following changes to the data and assumptions used in the
ECP analysis:

¦	Increased the indirect cost for engineering from $200,000 to $255,000 for the 50 MW
turbine. ECP's documentation indicated that this cost (as well as most of the other
direct installation and indirect costs) would be the same as for an SCR system on the
same turbine. The reported cost of $200,000 was inconsistent with this statement.

¦	Increased the contingencies cost for the 50 MW turbine from $76,486 to $244,101.
This change makes the cost consistent with ECP's statement that the cost for
contingencies is estimated to be equal to 5 percent of the total purchased equipment
cost, excluding the cost of the precious metals in the catalyst, sales taxes, and freight.

¦	Added a cost for the performance loss due to back pressure from the EMx system for
both turbines. ECP estimated the loss to be 0.5 percent, which is consistent with the
estimate in the 1993 ACT for SCR and the estimate OSEC used in a cost analysis for
SCONOx (EPA, 1993; OSEC, 1999). However, the ECP analysis did not include a
corresponding dollar amount for this element.

3-5


-------
¦	Changed the operating hours from 7,884 hr/yr to 8,000 hr/yr. This change also had a
small effect on the annual costs for utilities.

¦	Added costs for natural gas to generate steam for the 50 MW turbine using the same
procedures presented in the ECP analysis for the 180 MW turbine. ECP did not report
the basis for the amount of steam needed for the 180 MW turbine. Therefore, we
plotted the reported steam consumption versus turbine size for this unit and for two
turbines identified in a CARB analysis (CARB, 2004). We calculated the quantity of
steam needed for EMx on the 50 MW turbine using the regression equation from this
plot. Note that the unit cost for natural gas is $9.75/1000 scf. This was a reasonable
annual average cost in 2008, but it would be much too high for an analysis in 2014.

¦	Deleted the credit for recovery of precious metals in the spent catalyst because the
cost for replacement catalyst considers only the difference between the total purchase
price minus the value of the recovered material.

¦	Estimated the annualized cost of replacement catalyst (both the non-precious metal
substrate and the precious metal coating) using the future worth factor, whereas the
cost in the ECP analysis was the purchased cost divided by the 10-year replacement
interval.

¦	Estimated the cost of annual catalyst cleaning based on the average if data reported by
CARB (CARB, 2004) plus the amounts reported by ECP. Although ECP reported a
slightly higher cleaning cost for the 180 M W turbine than for the 50 MW turbine, an
analysis of all the cleaning data showed no correlation with turbine size. Thus, we
used the average of all reported costs for both turbines.

¦	Revised the indirect annual cost for administrative charges. ECP estimated that these
costs are the same as for an SCR system on the same turbines. We factored the cost as
2 percent of the TCI for the applicable EMx systems, which is consistent with the
approach for all control devices in the EPA Control Cost Manual. This resulted in
slightly higher costs.

¦	Increased the indirect costs for insurance, property tax, and capital recovery for both
turbines because the ECP analysis excluded the precious metal costs from the TCI
used in these calculations.

¦	Calculated capital recovery using an interest rate of 7 percent instead of 10 percent.

The capital costs for water injection in the combined EMx/water injection control
measure were estimated in 1999 dollars using the regression equation for the water injection
control measure (see Section 3.3.1) and then scaled to 2008 dollars using the Chemical
Engineering Plant Cost Index (CEPCI). Total annual costs for water injection were first
estimated in 1999 dollars using the regression equation for the water injection control option. On
average, 25 percent of these costs were estimated to be for indirect costs that are factored from

3-6


-------
the system capital cost, and the remaining 75 percent is for direct annual costs and overhead. To
estimate the total annual costs for water injection in 2008, the indirect costs were scaled from the
1999 estimate using the CEPCI, and the direct annual costs and overhead were assumed to be the
same as in 1999.

Table 3-2 summarizes the recommended cost effectiveness and capital to annual cost
ratios for implementing the EMx plus water injection NOx control measure. With an outlet
concentration of 2 ppmvd, this control measure achieves an average reduction ofpercent
relative to uncontrolled conventional diffusion combustion.

Table 3-2. Summary of Cost Effectiveness and Supporting Data for EMx Plus Water
Injection

Uncontrolled NOx

Emissions	Incremental Cost

	 EMx Outlet	Cost Capital to Relative to

Turbine Output, Cost Avg	Concentration, Effectiveness, Annual RACT Baseline

MW	Year ppmvd tpy	ppmvd	S/ton NOx Cost Ratio of WI, $/ton NOx

Large (50-180) 2008 160a >365	2.0	2.760	3.1	6,810

"Uncontrolled concentrations were not reported in the referenced analysis. Thus, the value used in this analysis is an
assumed average that results in the estimated 84 percent reduction for DLN combustion, as described in Section
3.3.3 of this report.

Based on regression of the data in the analysis, the best fit trend lines are represented by
the following power equations for the uncontrolled scenario (the R2 =1.0 for both equations
because there were only two data points in the analysis):

Total capital investment (2008 dollars) = 196928 x (MMBtu/hr) A 0 68

Total annual cost (2008 dollars) = 18747 x (MMBtu/hr) A 0 86

Best fit equations for incremental EMx costs relative to a RACT baseline of water
injection are:

Total capital investment (2008 dollars) = 156349 x (MMBtu/hr) A 0 68
Total annual cost (2008 dollars) = 17252 x (MMBtu/hr) A °-80

3-7


-------
3.2.3 EMx and Dry Low NOx Combustion; Gas Turbines—Natural Gas (NEMXDGTNG)

Table 3-3 summarizes the recommended cost effectiveness and capital to annual cost
ratios for implementing the EMx plus dry low NOx combustion control measure. With an outlet
concentration of 2 ppmvd, this control measure achieves an average reduction of 99 percent
relative to uncontrolled conventional diffusion combustion. For the same reasons noted in
Section 3.2.2, we recommend that this control measure be listed as an emerging technology in
the CMDB.

Table 3-3. Summary of Cost Effectiveness and Supporting Data for E]\lx Plus Dry Low
NOx Combustion

Uncontrolled	Incremental

NOx Emissions	Capital to Cost Relative to

Turbine Output,
MW

Cost
Year

Avg
ppmvd

tpy

EMx Outlet
Concentration,
ppmvd

Cost
Effectiveness,
S/ton NOx

Annual
Cost
Ratio

RACT Baseline
of DLN, $/ton
NOx

Small (4.2)

1999

134

<365

2.0

2.860

3.9

14,940

Small (23)

1999

174

>365

2.0

1.720

4.1

10,270

Large (170)

1999

210

>365

: (i

840

3.9

6,600

Large (50-180)

2008

o

>365

: (i

2,050

4.1

12,390

"Uncontrolled concentrations were not reported in the referenced analysis. Thus, the value used in this analysis is an
assumed average that results in the estimated 84 percent reduction for DLN combustion, as described in Section
3.3.3 of this report.

The recommended costs for EMx in 2008 dollars for the combined EMx/dry low NOx
combustion control measure are the same as in the estimate for the EMx/water injection control
measure described in Section 3.2.2. The recommended costs for EMx in 1999 dollars are based
on an analysis prepared by Onsite Sycom Energy Corporation (OSEC, 1999). For this analysis
the only changes we made to OSEC's analysis were to reduce the operating hours from
8,400 hr/yr to 8,000 hr/yr, which slightly reduced the energy penalty and utilities costs, and we
calculated the capital recovery factor using an interest rate of 7 percent instead of 10 percent.
Note that the total annual costs for natural gas (or purchased steam) are considerably lower in
this analysis than in the 2008 analysis because the unit cost of natural gas was considerably
lower in 1999.

The recommended total capital investment and total annual cost for dry low NOx
combustion in 1999 dollars for the combined EMx/dry low NOx combustion control measure are
the same as in the estimate for the dry low NOx combustion control measure alone as described
in Section 3.3.3. The recommended total capital investment for dry low NOx combustion in 2008

3-8


-------
dollars was estimated in 1999 dollars using the regression equation for the water injection control
measure (see Section 3.3.1) and then scaled to 2008 dollars using the CEPCI. The recommended
total annual costs for dry low NOx combustion consist of capital recovery plus the cost for parts
and repair; capital recovery costs in 2008 dollars were estimated by escalating the 1999 costs
using the CEPCI, and annual parts and repairs costs were assumed to be the same in 2008 as in
1999.

Based on regression of the data in both the 1999 and 2008 cost analyses, the Ix-sl fit trend
lines are represented by the following power equations for the uncontrolled scenario (the R2 =1.0
for the equations in 2008 dollars because there were only two data points in the analysis; R2 for
the equations in 1999 dollars round to 1.0 when only two significant figures are presented):

Total capital investment (1999 dollars) = 58237 x (MMBtu/hr) A 0 78

Total annual cost (1999 dollars) = 15004 x (MMBtu/hr) A 0 78
Total capital investment (2008 dollars) 120S92 x (MMBtu/hr) A 074

Total annual cost (2008 dollars) = 20041 x (MMBtu/hr) A °-80

Best fit equations for incremental EMx costs relative to a RACT baseline of DLN
combustion are:

Total capital investment (1999 dollars) = 65163 x (MMBtu/hr) A 0 72

Total annual cost (1999 dollars) = 13702 x (MMBtu/hr) A 0 76

Total capital investment (2008 dollars) = 156349 x (MMBtu/hr) A 0 68

Total annual cost (2008 dollars) = 17252 x (MMBtu/hr) A °-80
3.3 Recommended Changes

This section presents updated cost estimates for combustion turbine control measures that
are currently in the CMDB, and it describes the basis for such changes. These changes include
both more recent costs for some control measures as well as minor revisions to existing estimates
for other control measures. The changes affect both cost per ton values and equations.

3-9


-------
This section also identifies applicable SCCs for the new control measures described in
Section 3.2, and it identifies additional SCCs for which the control measures in this section are
applicable.

3.3.1 Water Injection; Gas Turbines—Natural Gas (NWTINGTNG)

Recommended updates to the costs for water injection are based on analyses in a report
prepared by OnSite Sycom Energy Corporation for the U.S. Department of Energy (OSEC,
1999). OSEC estimated costs for some of the same small turbine model sizes as in EPA's 1993
ACT document (4 MW and 23 MW). OSEC obtained water injection equipment costs in 1999
dollars. They then estimated total capital investment and total annual costs using the same
procedures as in the 1993 ACT document, and they concluded that 1999 costs for water injection
were essentially the same as the 1990 costs presented in the ACT document. Because the ACT
analysis included a greater number of models over a wider range of sizes, RTI recommends
continuing to use the cost data from the ACT analysis in the CMDB, except the cost year should
be updated from 1990 to 1999. RTI also recommends the four additional changes noted below.

Our second recommendation is to split the record for small sources into two records—
one for sources with uncontrolled emissions less than 365 tpy, and the other for emissions greater
than 365 tpy. The 2006 AirControlNET Documentation Report indicates that small sources are
turbines with design outputs up to 34.4 MW. Four model turbines in the ACT analysis have
outputs below this threshold. The two turbines with uncontrolled emissions <365 tpy have an
average cost effectiveness of $ 1,790/ton of NOx. The two turbines with uncontrolled emissions
>365 tpy have an average cost efllvli veness of $ 1,000/ton of NOx.

Our third recommendation is to revise the control efficiency for water injection from 76
percent to 72 percent. The 76 percent control level is the average reduction for all 6 model
turbines in the 1993 ACT analysis. Five of those models were guaranteed to reduce NOx
emissions to less than 42 ppmvd, while the sixth was guaranteed to meet 25 ppmvd. Although
water injection may be more effective on some combustion turbines than others, 42 ppmvd is the
generally accepted threshold. Thus, we think this threshold should be incorporated in the CMDB.
The average reduction of the 5 models in the 1993 ACT analysis with an outlet concentration of
42 ppmvd was 72 percent.

Our fourth recommendation is to use a capital to annual cost ratio of 2.4 in the new
record for small sources with uncontrolled emissions >365 tpy; this is the average value for the
two turbines in the ACT analysis in this size range. (The capital to annual cost ratio for the small
sources with uncontrolled emissions <365 tpy would remain at 3.1 because this is the average

3-10


-------
value for the two turbines in this size range; it is not clear why this value was applied for all
small sources in the current version of the CMDB.) The total annual costs in this calculation are
based on using a 7 percent interest rate in the calculation of capital recovery, instead of the 10
percent value in the 1993 ACT. Even if capital recovery was estimated using the 10 percent
interest rate, it is not clear how the 3.1 value was developed.

Our fifth recommendation is to revise the constants in the CMDB table of equations for
estimating capital and annual costs. Based on regression of the data in the 1993 ACT, the best fit
trend lines are represented by the following revised power equations for both uncontrolled and
RACT baseline scenarios:

Total capital investment (1999 dollars) = 27665 x (MMBtu/hr) A °-()9 (R2 =0.97)

Total annual cost (1999 dollars) = 3700.2 x (MMBtu/hr) A °-9- (R2=0.95)
3.3.2 Steam Injection; Gas Turbines—Natural Gas (NSTINGTNG)

The only available information on the cost of steam injection was in the 1999 report from
Onsite Sycom Energy Corporation (OSEC, 1999). OSEC discussed steam injection only in the
context of large GE Frame 7F turbines (170 M W). They noted that only the first such model,
operational in 1990 when the ACT analysis was being conducted, was equipped with steam
injection. All subsequent units (at least through 1999) were equipped with DLN combustion
technology.

Because the limited available information suggests that steam injection costs, like water
injection costs, were essentially the same in 1999 as in 1990, we recommend continuing to base
the steam injection costs on the results in the 1993 ACT, but update the cost year from 1990 to
1999. In addition, as for water injection, we recommend splitting the one record for small
sources into two records—one for sources with uncontrolled NOx emissions <365 tpy, and the
other for uncontrolled NOx emissions >365 tpy. This split results in average cost effectiveness
values of $ 1,690/ton of NOx for the small sources with uncontrolled NOx emissions
<365 tons/yr and $820/ton of NOx for the small sources with uncontrolled NOx emissions
>365 tons/yr. The capital cost to annual cost ratios also are slightly less than the current values in
the CMDB.

Based on regression of the data in the 1993 ACT, the best fit trend lines are represented
by the following revised power equations for both uncontrolled and RACT baseline scenarios:

3-11


-------
Total capital investment (1999 dollars) = 43092 x (MMBtu/hr) A 0 82 (R2=0.95)

Total annual cost (1999 dollars) = 7282 x (MMBtu/hr) A 0 76 (R2=0.96)

3.3.3 Dry Low NOx Combustion; Gas Turbines—Natural Gas (NDLNCGTNG)

Dry low NOx (DLN) combustion technology premixes air and a lean fuel mixture that
significantly reduces peak flame temperature and thermal NOx formation. In some cases, this
can be accomplished by using low NOx burners, but in other cases, the combustor design itself
differs as well as the burner design. For example, the DLN combustor volume is typically twice
that of a conventional combustor (OSEC, 1999). Therefore, we recommend revising the current
control technology name in the CMDB from "Low NOx Burners" to "Dry Low NOx
Combustion." In addition, the CM abbreviation should be changed from NLNBUGTNGto
NDLNCGTNG.

Recommended updates to the costs for DLN Combustion are based on analyses in a
report prepared by Onsite Sycom Energy Corporation for the U.S. Department of Energy (OSEC,
1999). OSEC estimated costs for some of the six turbines with design outputs ranging from
4 MW to 169 MW.

OSEC obtained installed equipment costs and annual repair costs in 1999 dollars from
three turbine manufacturers, but there are some uncertainties in the data. Although the reported
tabular summary indicates the equipment costs are incremental relative to the cost of a
conventional combustor, the text of the report states that the costs for 169 MW turbines are the
total cost to replace a conventional combustor (which may explain why the regression equation
for the capital costs is linear rather than a power function). Annual costs for parts and repair for
some of the turbines were proprietary for two of the small turbines and thus could not be
reported. As a result, the annual costs for those turbines are biased low. In addition, because parts
and repair costs were unavailable for the 169 MW turbine, OSEC assumed these costs could be
represented by the costs for the 23 MW turbine.

The only change we made to the assumptions and data reported by OSEC was to
calculate capital recovery using an interest rate of 7 percent instead of 10 percent.

Table 3-4 summarizes the recommended new cost effectiveness and capital to annual cost
ratios for implementing the DLN combustion NOx control technology. In addition to changing
these costs in the CMDB, we also recommend changing the control efficiency for DLN
combustion applied to small sources from 68 percent to 84 percent. The 84 percent level is

3-12


-------
currently used for large sources, and it is consistent with the efficiency for DLN combustion (or
low NOx burners) in the 1993 ACT. It appears the 68 percent entry was a data transcription error
because that is the control efficiency for water injection applied to oil-fired turbines.

Table 3-4. Summary of Cost Effectiveness and Supporting Data for DLN Combustion





Uncontrolled NOx











Emissions



Outlet

Cost

Capital to

Turbine

Cost





- Concentration,

Effectiveness,

Annual Cost

Output, MW

Year

Avg. ppmvd

tpy

ppmvd

$/ton NOx

Ratio

Small (4-23)

1999

152

<365

25

300

5.0

Large (170)

1999

210

>365

25

130

7.4

Based on regression of the data in both analyses, the best fit trend lines are represented by
the following revised equations for both uncontrolled and RACT baseline scenarios:

Total capital investment (1999 dollars) = 2860.6 x (MMBtu/hr) + 25427 (R2=1.0)

Total annual cost (1999 dollars) = 584.5 x (MMBtu/hr) A 0 96 (R2=0.95)

3.3.4 SCR and Water Injection; Gas Turbines—Natural Gas (NSCRWGTNG)

Recommended updates to the costs for SCR combined with water injection are based on
two sets of cost analyses. One set of costs is in 1999 dollars for three turbines ranging in size
from 4.2 MW to 161 MW (OSEC, 1999). The second is in 2008 dollars for two larger turbines
with design outputs of 50 MW and 180 MW (ECP, 2008). For SCR, the referenced analyses
estimated direct installation costs and indirect costs based on scaling from the purchased
equipment costs using standard factors as in the Control Cost Manual. Annual costs were
estimated for the same cost elements that were used in the SCR analysis in the 1993 ACT. Water
injection costs for the two smallest turbines in the 1999 analysis were estimated as described
above for the water injection control option. Water injection costs for the large turbines were not
estimated in the referenced analyses.

For the purposes of developing 1999 cost inputs for the CMDB, we made the following
changes to the data and assumptions used in the OSEC analysis:

¦ Increased the engineering cost for SCR for the 161 MW turbine from $100,000 to
$228,865. The revised value is equal to 10 percent of the purchased equipment cost,
which is consistent with the approach used for the smaller turbines. The report did not
explain why $100,000 was used instead of the factor.

3-13


-------
¦	Estimated performance penalty costs and electricity costs for the blower and pumps in
the ammonia injection system using operating hours of 8,000 hr/yr instead of

8,400 hr/yr.

¦	Calculated capital recovery for the SCR system using an interest rate of 7 percent
instead of 10 percent.

¦	Calculated annual catalyst replacement and disposal costs using a future worth factor
instead of a capital recovery factor.

¦	Estimated total capital investment and total annual costs for the lc-> I \I\V lurhine
using the regression equations for the water injection control option. (Maybe it would
be better to drop the large model from this analysis and just present 1999 costs for
small turbines and 2008 costs for large turbines.)

For the purposes of developing 2008 cost inputs for the CMDB, we started with the ECP
analysis for SCR costs and then made the following changes to the data and assumptions:

¦	Calculated the performance penalty for SCR using an electricity cost of $0.06/kwh
instead of $0.1/kwh and 8,000 hr/yr instead of 8,400 hr/yr. In addition, although it
appears that the referenced analysis assumed a performance loss equal to 0.5 percent
of the turbine's design output, the cited cost was significantly greater than it should
be for that percentage loss, even if the cited electricity cost and operating hours were
used in the calculation. We changed the cost to be consistent with the calculated
amount.

¦	Calculated capital recovery for the SCR system using an interest rate of 7 percent
instead of 10 percent.

¦	Estimated capital costs for water injection in 1999 dollars using the regression
equation for the water injection control option, and then scaled the costs to 2008
dollars using the CEPCI.

¦	Estimated total annual costs for water injection following the same procedure
described in Section 3.2.2 for the water injection portion of a combined water
injection and EMx control measure. Thus, the total annual costs for water injection
are the same in both control measures.

Table 3-5 summarizes the recommended new cost effectiveness and capital to annual cost
ratios values for implementing SCR plus water injection on natural gas-fired combustion
turbines. Table 3-5 also presents revised incremental costs of SCR relative to a RACT baseline
of water injection for the different categories of turbines. Note that the SCR outlet NOx level
was assumed to be 2.5 ppmvd in the ECP analysis, which results in an overall control efficiency
of 98 percent versus the 94 percent for the OSEC and ACT analyses.

3-14


-------
Table 3-5. Summary of Cost Effectiveness and Supporting Data for SCR Plus Water
Injection

Turbine
Output, MW

Cost
Year

Uncontrolled NOx
Emissions

SCR Outlet
Concentration,
ppmvd

Cost
Effectiveness,
$/ton NOx

Capital to
Annual
Cost Ratio

Incremental Cost
Relative to RACT
Baseline of WI,
$/ton NOx

Avg. ppmvd

tpy

Small (4.2)

1999

134

<365

9

2,790

3.0

5,840

Small (22.7)

1999

174

>365

9

1,370

2.9

3,130

Large (161)

1999

210

>365

9

1,070

1.5

l.ii'Ni

Large (50-180)

2008

o

>365

2.5

1,830

2.7

3.170

"Uncontrolled concentrations were not reported in the referenced analysis. Thus, the value used in this analysis is an
assumed average that results in the estimated 84 percent reduction for DLN combustion, as described in Section
3.3.3 of this report.

Based on regression of the data in both analyses, the revised best fit trend lines are
represented by the following power equations for both uncontrolled scenarios (R2=l for the 2008
costs because the analysis was based on only two data points):

Total capital investment (1999 dollars) = 62962 x (MMBtu/hr) A 0 ()6 (R2=1.0)

Total annual cost (1999 dollars) = 8590 x (MMBtu/hr) A 0 87 (R2=0.99)

Total capital investment (2008 dollars) = 34533 x (MMBtu/hr) A 0 85 (R2=1.0)

Total annual cost (2008 dollars) = 6794 x (MMBtu/hr) A 0 94 (R2=1.0)

Revised best fit equations for incremental SCR costs relative to a RACT baseline of
water injection are (R2=l for the 2008 costs because the analysis was based on only two data
points).

Total capital investment (1999 dollars) = 37193 x (MMBtu/hr) A 0 63 (R2=1.0)

Total annual cost (1999 dollars) = 12065 x (MMBtu/hr) A 0 64 (R2=1.0)

Total capital investment (2008 dollars) = 10323 x (MMBtu/hr) A 0 96 (R2=1.0)

Total annual cost (2008 dollars) = 3106.1 x (MMBtu/hr) A 0 94 (R2=1.0)

3-15


-------
3.3.5 SCR and Steam Injection; Gas Turbines—Natural Gas (NSCTSGTNG)

Combined costs for SCR and steam injection were not presented in any available
references. Thus, costs for combined control systems were estimated in 1999 dollars for four
model turbines ranging from 4 MW to 161 MW using the procedures described above for steam
injection alone and for SCR as part of combined SCR and water injection control systems.
Specifically, steam injection costs for each model turbine were assumed to be the same as in the
1993 ACT, consistent with the description above for steam injection control costs. Since OSEC
did not estimate SCR costs for the specific turbines in this analysis, we estimated the SCR costs
using the trendlines that we developed for incremental SCR costs relative to a RACT baseline of
water injection. We then summed the separate SCR and steam injection costs to obtain the
combined system costs.

We also estimated costs for a combined steam injection and SCR control measure in 2008
dollars. The SCR portion of the costs are the same as for SCR in the combined water injection
plus SCR control measure, as described in Section 3.3.4. Total capital investment for the steam
injection portion were estimated in 1999 dollars using the regression equation developed for
steam injection alone, as described in Section 3.3.2. These costs were escalated to 2008 costs
using the CEPCI. Total annual costs for steam injection were first estimated in 1999 dollars
using the regression equation for the steam injection control option (see Section 3.3.2). On
average, 40 percent of these costs were estimated to be for indirect costs that are factored from
the system capital cost, and the remaining 60 percent is for direct annual costs and overhead. To
estimate the total annual costs for steam injection in 2008, the indirect costs were scaled from the
1999 estimate using the CF.PCI, and the direct annual costs and overhead were assumed to be the
same as in 1999.

Table 3-6 summarizes the recommended new cost effectiveness and capital to annual cost
ratios values for implementing SCR plus steam injection on natural gas-fired combustion
turbines. Table 3-6 also presents revised incremental costs of SCR relative to a RACT baseline
of steam injection for the different categories of turbines. Note that the incremental costs are
slightly different from the costs in Table 3-5. The costs should be the same for a given turbine
category. They differ because the two analyses examined a different number of turbines, and the
sizes were not exactly the same. At a later date, the analysis could be improved by combining the
SCR costs from both analyses and developing a single set of incremental SCR costs.

3-16


-------
Table 3-6. Summary of Cost Effectiveness and Supporting Data for SCR Plus Steam
Injection (SI)

Turbine Output,
MW

Cost
Year

Uncontrolled NOx
Emissions

SCR Outlet
Concentration,
ppmvd

Cost
Effectiveness,
$/ton NOx

Capital to
Annual
Cost Ratio

Incremental Cost
Relative to RACT
Baseline of SI,
$/ton NOx

Avg. ppmvd

tpy

Small (4.2)

1999

155

<365

9

2,570

3.3

5,550

Small (26.8)

1999

142

>365

9

1,380

3.1

2,870

Large (83-161)

1999

300

>365

9

570

2.7

1 .S III

Large (50-180)

2008

o

>365

2.5

1,420

3.9

3.170

"Uncontrolled concentrations were not reported in the referenced analysis. Thus, the value used in this analysis is an
assumed average that results in the estimated 84 percent reduction for DLN combustion, as described in Section
3.3.3 of this report.

Based on regression of the data in the analysis, the revised best fit trend lines are
represented by the following power equations for the uncontrolled scenario (R2= 1 for the 2008
costs because the analysis was based on only two data points):

Total capital investment (1999 dollars) = 72169 x (MMBtu/hr) A 0 ()() (R2=0.99)

Total annual cost (1999 dollars) = 1755 1 x (MMBtu/hr) A 0 72 (R2=0.98)

Total capital investment (2008 dollars) = 46492 x (MMBtu/hr) A 0 82 (R2=1.0)

Total annual cost (2008 dollars) = 8704 x (MMBtu/hr) A 0 86 (R2=1.0)

Revised best fit equations for incremental SCR costs relative to a RACT baseline of
steam injection are assumed to be the same as noted above in the discussion of costs for SCR and
water injection.

3.3.6 SCR and Dry Low NOx Combustion; Gas Turbines—Natural Gas (NSCRDGTNG)

Updated costs for combined SCR and DLN combustion control systems were estimated
in 1999 dollars for all turbine sizes, 2007 dollars for small turbines, and 2008 dollars for large
turbines. The 1999 costs were estimated by combining the separate costs for DLN combustion
and SCR provided by Onsite Sycom Energy Systems (OSEC, 1999). The 2007 costs were
estimated by combining SCR costs developed by Energy and Environmental Analysis in a report
prepared for EPA with the OSEC costs for DLN combustion in 1999 dollars, escalated to 2007
dollars (EEA, 2008). Similarly, costs in 2008 dollars were estimated by combining SCR costs

3-17


-------
developed by EmeraChem Power in an analysis for the Bay Area Air Quality Management
District with escalated DLN combustion costs (ECP, 2008). The EEA analysis provided only
capital costs; therefore, we estimated annual costs using the same factors provided in ECP's
analysis of costs in 2008 dollars. For both the 2007 and 2008 cost estimates, DLN capital costs
and capital recovery were escalated from 1999 dollars using the CEPCI, and annual parts and
repairs costs were assumed to be the same in all three years.

Table 3-7 summarizes the recommended new cost effectiveness and capital to annual cost
ratios values for implementing SCR plus dry low NOx combustion on natural gas-fired
combustion turbines. Table 3-7 also presents revised incremental costs of SCR relative to a
RACT baseline of steam injection for the different categories of turbines. Note that the SCR
outlet NOx level was assumed to be 2.5 ppmvd in the ECP analysis, which results in an overall
control efficiency of 98 percent versus the 94 percent for the OSEC analyses. We also used an
outlet concentration of 2.5 ppmvd to estimate emissions to use with EEA's 2007 costs. The ECP
and EEA analyses did not specify inlet NOx emissions concentrations to the SCR; therefore, we
assumed 25 ppmvd, as in other DLN analyses. We also assumed an average uncontrolled
emissions level of 160 ppmvd for all models so that the overall control efficiency of the DLN
combustion plus the SCR was 98 percent. Note that the incremental costs in 1999 dollars are
significantly higher than those for SCR following water injection and steam injection; this is due
to the inlet concentration being 25 ppmvd for this analysis and 42 ppmvd for water injection and
steam injection.

Table 3-7. Summary of Cost Effectiveness and Supporting Data for SCR Plus DLN
Combustion

Turbine Output.
MW

Cost
Year

Uneontrolled NOx
Emissions

SCR Outlet
Concentration,
ppmvd

Cost
Effectiveness,
$/ton NOx

Capital to
Annual
Cost Ratio

Incremental Cost
Relative to RACT
Baseline of DLN,
$/ton NOx

Av& ppmvd

tpy

Small (4.2)

1999

134

<365

9

1,800

2.9

11,900

Small (26.8)

1999

174

>365

9

990

3.6

6,320

Large (161)

1999

210

>365

9

390

4.2

3,340

Small (1-10.2)

2007

o

<365

2.5

2,910

4.3

18,900

Small (25)

2007

o

>365

2.5

1,460

3.8

7,510

Large (50-180)

2008

o

>365

2.5

1,040

4.5

5,560

"Uncontrolled concentrations were not reported in the referenced analysis. Thus, the value used in this analysis is an
assumed average that results in the estimated 84 percent reduction for DLN combustion, as described in Section
3.3.3 of this report.

3-18


-------
Based on regression of the data in each analysis, the best fit trend lines are represented by
the following power equations for uncontrolled scenarios (R2=l for the 2008 costs because the
analysis was based on only two data points, and note that the R2 for the 2007 equations is not
meaningful because the DLN portion of the costs are based on a regression equation instead of
independent, model-specific data):

Total capital investment (1999 dollars) = 24854 x (MMBtu/hr) A 0 79 (R2=l .0)

Total annual cost (1999 dollars) = 12725 x (MMBtu/hr) A 0 69 (R2=l .0)

Total capital investment (2007 dollars) = 187647 x (MMBtu/hr) A 0 54 (R2=l .0)

Total annual cost (2007 dollars) = 2782 x (MMBtu/hr) + 167494 (R2=1.0)

Total capital investment (2008 dollars) = 14790 x (MMBtu/hr) A 0 97 (R2=1.0)

Total annual cost (2008 dollars) = 5263.5 x (MMBtu/hr) A °-90 (R2=1.0)

The equations to estimate incremental costs for SCR relative to a RACT baseline of dry
low NOx combustion in 1999 dollars and 2008 dollars are assumed to be the same as noted in
Section 3.3.4 for incremental costs relative to a RACT baseline of water injection. Incremental
costs for SCR relative to a RACT baseline of water injection in 2007 dollars are estimated using
the following equations'

Total capital imesiment (2007 dollars) = 210883 x (MMBtu/hr) A 046 (R2=1.0)

Total annual cost (2007 dollars) = 1894 x (MMBtu/hr) + 185570 (R2=0.99)
3.3.7 Water Injection; Gas Turbines—Oil (NWTINGTOL)

No new data are available on costs of water injection for oil-fired combustion turbines.
However, because the water injection costs for natural gas-fired turbines were determined to be
essentially the same in 1999 as in 1990, we assume the same would be true for water injection on
oil-fired turbines; the costs for both types of turbines also were the same in the 1993 ACT
analysis. Therefore, we recommend continuing to base costs on the results of the 1993 ACT
analysis, but to update the cost year from 1990 to 1999. In addition, we changed the size of the
large model in the ACT analysis from 83.3 MW to 84.7 MW because it appears the incorrect

3-19


-------
model was used in the ACT analysis. As for the natural gas-fired turbines, we also recommend
splitting the single record for small sources into two records—one for source with uncontrolled
NOx emissions <365 tpy, and the other for sources with uncontrolled NOx emissions >365 tpy.
The resulting cost effectiveness values for the turbines with uncontrolled NOx emissions
<365 tpy and >365 tpy are $l,630/ton of NOx and $960/ton of NOx, respectively. The capital to
annual cost ratios also change slightly.

As for other control technologies, the constants in the equations to estimate total capital
costs and total annual costs differ from those in the regression analyses performed in Excel. In
this case, the differences are small, but we recommend revising the constants so that all
equations are developed based on the same approach. The revised equations for both the
uncontrolled and RACT baseline scenarios are:

Total capital investment (1999 dollars) = 43255 x (MMBtu/hr) A 0 ()0 (R2=1.0)

Total annual cost (1999 dollars) = 6796.8 x (MMBtu/hr) A °-80(R2=1.0)

3.3.8 SCR and Water Injection; Gas Turbines—Oil (NSCRWGTNG)

SCR costs were developed in a BACT analysis for a 48 MW oil-fired combustion turbine
(FMPA, 2004). Because water injection costs in 2004 dollars are not available, we calculated
costs in 1999 dollars as described above for the water injection option, and then estimated costs
in 2004 dollars by scaling up the 1999 capital costs (and capital recovery) using the CEPCI;
other annual operating and maintenance costs were assumed to be unchanged. We used the SCR
capital cost as presented in the FMPA analysis, but we made several changes to the annual costs.
Although the original values may have been appropriate for the specific application evaluated by
FMPA, the following changes were made to be consistent with the calculations for other controls
in this analysis:

¦	Estimated O&M costs assuming operation for 8,000 hr/yr instead of 4,422 hr/yr.

¦	Excluded cost for one week of lost power generation while catalyst is being replaced,
assuming that catalyst replacement can be performed during scheduled annual
downtime.

¦	Reduced sales tax and freight cost for catalyst from 12.25 percent of the purchased
cost to 8 percent of the purchased cost.

¦	Deleted capital recovery cost for catalyst because the catalyst is replaced annually.

3-20


-------
¦	The reported annual cost for ammonia was based on a stoichiometric ratio of 1.4
(possibly because they assumed a significant generation of N02 relative to NO).
They also applied a factor of 1.05, apparently to account for ammonia slip, as in the
Control Cost Manual procedures for SCR on boilers. However, both factors should
not be needed. For this analysis, we used just the 1.05 factor (also used the reported
unit cost of $750/ton of ammonia, which may have been high for 2004).

¦	Reduced the property tax factor from 2.75 percent of the TCI to 1 percent of the TCI.

Table 3-8 summarizes the recommended cost effectiveness and capital lo minimi cost
ratios values for implementing SCR plus water injection on oil-fired combustion turbines.

Table 3-8 also presents incremental costs of SCR relative to a RACT baseline of water injection.
The 1990 costs are essentially the same as the costs currently in the CMDB, except that we
recommend splitting the one record for small sources into two records.

Table 3-8. Summary of Cost Effectiveness and Supporting Data for SCR Plus Water
Injection (WI) for Oil-Fired Turbines





Uncontrolled NOx





Incremental Cost





Emissions

SCR Outlet Cost

Capital to

Relative to RACT

Turbine

Cost





" Concentration. Effectiveness,

Annual

Baseline of WI,

Output, MW

Year

Avg. ppmvd

tpy

ppnml S/tonNOx

Cost Ratio

$/ton NOx

Small (3.3)

1990

179

<365

IS 3,200

2.9

7,620

Small (26.3)

1990

211

>365

IS 1,320

2.3

2,450

Large (84)

1990

228

>365

18 1,000

2.4

2,210

Large (48)

2004

200"

>365

5 1,560

2.3

4,790

aThe referenced analysis did nol report an uncontrolled emissions level. The value used in this analysis is the
average of the uncontrolled emissions concentrations for oil-fired model turbines in the 1993 ACT.

Based on regression of the data in the 1993 ACT, the best fit trend lines are represented
by the following revised power equations for the uncontrolled scenario:

Total capital investment (1990 dollars) = 95837 x (MMBtu/hr) A 0 62 (R2=0.99)

Total annual cost (1990 dollars) = 25990 x (MMBtu/hr) A °-70 (R2=1.0)

Revised best fit equations for incremental SCR costs relative to a RACT baseline of
water injection are:

Total capital investment (1990 dollars) = 4744 x (MMBtu/hr) + 368162 (R2=1.0)

3-21


-------
Total annual cost (1990 dollars) = 1522.5 x (MMBtu/hr) + 142643 (R2=1.0)

We could not develop equations for this control system in 2004 dollars because 2004 data
are available for only one turbine, and thus are insufficient for this purpose.

3.3.9	Water Injection; Gas Turbines—Jet Fuel (NWTINGTJF)

The current CMDB assumes costs for jet fuel-fired turbines are the same as for oil-fired
turbines. Thus, we recommend the same changes for jet fuel fired turbines as noted al">o\ e lor oil-
fired turbines.

3.3.10	SCR and Water Injection; Gas Turbines—Jet Fuel (NSCTWGTJF)

The current CMDB assumes costs for jet fuel-fired turbines are the same as for oil-fired
turbines. Thus, we recommend the same changes for jet fuel fired turbines as noted above for oil-
fired turbines.

3.3.11	Applicable Control Measures for Gas Turbine SCCs

The first column in Table 3-9 lists all of the gas turbine SCCs that are associated with one
or more gas turbine control measures in the CMDIi luMe called "Table 03_SCCs." In addition,
the last seven SCCs in Table 3-9 are additional gas turbine SCCs that are not currently assigned
any NOx control measures in the CMDB. These seven SCCs, as well as many of the others at the
top of Table 3-9, were identified with NOx emissions in an EPA query of the NEI for facilities in
the Ozone Transport Group Assessment Region (i.e., 37 states that are partially or completely to
the east of 100°W longitude). The first 1 1 control measures in column headings in Table 3-9 are
the gas turbine control measures that are currently in the CMDB; the last three column headings
are the new control measures identified in this review and described in Section 3.2 of this report.

Each control measure that was determined to be applicable for a specific SCC is
identified by either an "E" or an "N" in the cell at the intersection of the applicable SCC row and
the control measure column. An "E" means the control measure is already listed in the CMDB
for the particular SCC, and we concur with that designation. An "N" means the control measure
is not currently linked to a particular SCC, but we recommend adding this link in the database. In
some cases, we recommend applying new links between existing control measures and existing
SCCs. For example, some of the SCCs are for turbines that are fired with relatively uncommon
fuels such as landfill gas or gasoline. We have not located any analyses that determined the
applicable controls and related costs for gas turbines fired with such fuels. In order to conduct
CoST modeling analyses for these turbines, the most representative available control measures

3-22


-------
Table 3-9. Recommended Control Measures for Gas Turbine SCCs

Applicable Gas Turbine Control Measures for the SCCd



see

see





hJ
O

H
O

z

-J
O

H
O
£

li

H
O

Z

li

O
£

O

z

H
O

z

STTNGTNG

O
Z
H
O
U

O
Z
H
O

O
Z
H
O

C/2

O
Z
H
O

h
C
<
h

O

z

H
O
U

O
Z
H
O

O
Z
H
O
Q

scca

Level
lb

Level
T

SCC Level 3

SCC Level 4

HH

H
£

C£

U
cn

HH

H
£

C£

U
cn

HH

H
£

Z
-J
Q

C£

U
cn

C£

U
cn

U
cn

e

H
U

|

U



Z

Z

z

Z

z

Z

Z

Z

Z

Z



z

z

Z

20200101

ICE

Ind

Distillate Oil (Diesel)

Turbine

E

E

























20200103

ICE

Ind

Distillate Oil (Diesel)

Turbine: Cogeneration

E

E

























20200108

ICE

Ind

Distillate Oil (Diesel)

Turbine: Evap Losses

D

D

























20200109

ICE

Ind

Distillate Oil (Diesel)

Turbine: Exhaust

E

E

























20200201

ICE

Ind

Natural Gas

Turbine









E

E

E

E

E

E

D

N

N

N

20200203

ICE

Ind

Natural Gas

Turbine: Cogeneration









E

E

E

E

E

E

D

N

N

N

20200208

ICE

Ind

Natural Gas

Turbine: Evap Losses









D

D

D

D

D

D

D







20200209

ICE

Ind

Natural Gas

Turbine: Exhaust









E

E

E

E

E

E

D

N

N

N

20200701

ICE

Ind

Process Gas

Turbine









Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20200705

ICE

Ind

Process Gas

Refinery Gas: Turbine









Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20200713

ICE

Ind

Process Gas

Turbine: Evap Losses





















D







20200714

ICE

Ind

Process Gas

Turbine: Exhaust









Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20200901

ICE

Ind

Kerosene/Naphtha (Jet
Fuel)

Turbine





E

E





















20200908

ICE

Ind

Kerosene/Naphtha (Jet
Fuel)

Turbine: Evap Losses





D

D





















20200909

ICE

Ind

Kerosene/Naphtha (Jet
Fuel)

Turbine: Exhaust





E

E





















20201008 ICE

Ind	Liquified Petroleum Gas

(LPG)

Turbine: Evap Losses

D

(continued)


-------
Table 3-9. Recommended Control Measures for Gas Turbine SCCs (continued)

Applicable Gas Turbine Control Measures for the SCCd

SCC SCC
Level Level
SCCa lb	2C

SCC Level 3

SCC Level 4

-J
O

H

O

z
—

H
£

-j
o

H

o

%

U
cn

H
O

Z

HH

H
£

I
I

U

O

z

H
O

z

HH

H
£

O
z

H

o

z

HH

H

m

O
Z
H
O
U
Z
-J
Q

O
z

H

0

1

U
in

O
Z
H
O

in
Pi

u

CA)

O

z

H
O

U

CA)

h

c
<
h

c

o

z

H

o

U
H

U

O

z

H
O

o

z

H

o

Q



20201009

ICE

Ind

Liquified Petroleum Gas
(LPG)

Turbine: Exhaust

Ne

Ne













D







20201011

ICE

Ind

Liquified Petroleum Gas
(LPG)

Turbine

Ne

Ne













D







20201013

ICE

Ind

Liquified Petroleum Gas
(LPG)

Turbine: Cogeneration

Ne

\













D







20300102

ICE

C/l

Distillate Oil (Diesel)

Turbine

E

E





















20300108

ICE

C/l

Distillate Oil (Diesel)

Turbine: Evap Losses

D

D





















20300109

ICE

C/l

Distillate Oil (Diesel)

Turbine: Exhaust

E

E





















20300202

ICE

C/l

Natural Gas

Turbine





E

E

E

E

E

E

D

N

N

N

20300203

ICE

C/l

Natural Gas

Turbine: Cogeneration





E

E

E

E

E

E

D

N

N

N

20300208

ICE

C/l

Natural Gas

Turbine: Evap Losses





D

D

D

D

D

D

D







20300209

ICE

C/l

Natural Gas

Turbine: Exhaust





E

E

E

E

E

E

D

N

N

N

20300701

ICE

C/l

Digester Gas

Turbine





Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20300708

ICE

C/l

Digester Gas

Turbine: Evap Losses

















D







20300709

ICE

C/l

Digester Gas

Turbine: Exhaust





Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20300801

ICE

C/l

Landfill Gas

Turbine





Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20300808

ICE

C/l

Landfill Gas

Turbine: Evap Losses

















D







20300809

ICE

C/l

Landfill Gas

Turbine: Exhaust





Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

20400301

ICE

ET

Turbine

Natural Gas





N

N

N

N

N

N

D

N

N

N

20400304

ICE

ET

Turbine

Landfill Gas





Ne

Ne

Ne

Ne

Ne

Ne

D

Ne

Ne

Ne

(continued)


-------
Table 3-9. Recommended Control Measures for Gas Turbine SCCs (continued)

Applicable Gas Turbine Control Measures for the SCCd

scca

see

Level
lb

see

Level
2C

SCC Level 3

SCC Level 4

-J

o

H

o

z

HH

H

-J

o

H

o

%
U
/
y

to
H

o

z
—

H

to

I

u
/.

O

z

H
O

z

HH

H

O
z

H

o

z

HH

H

o

Z
H
O
U
Z
-J
Q

O
z

H

0

1

U

o

z

H
O

in
Pi

u

CA)

O

z

H
O

U
in

h
C
<
h

e

o

z

H
O
U
H

U

O

z

H
O

o

z

H

o

Q



50100420

WD

SWD-G

Landfill Dump

Waste Gas Recovery:
GT







\

Ne Ne Ne Ne Ne D Ne Ne Ne

20201609

ICE

Ind

Methanol

Turbine: Exhaust











20201701

ICE

Ind

Gasoline

Turbine











20300901

ICE

C/I

Kerosene/Naphtha (Jet
Fuel)

Turbine: JP-4





N

N



20400302

ICE

ET

Turbine

Diesel/Kerosene

N

N







20400303

ICE

ET

Turbine

Distillate Oil

N

N







20400305

ICE

ET

Turbine

Kerosene/Naphtha





N

N



20400399

ICE

ET

Turbine

Other Not Classifiedf

Ne

Ne







aSCCs in regular font are associated with one or more gas turbine control measures in the current CMDB. The SCCs in bold font represent gas turbine activities
that were identified with NOx emissions in the Ozone Transport Assessment Group Region analysis but are not associated with gas turbine control measures in
the current CMDB.

bICE means "Internal Combustion Engines" and WD means "Waste Disposal."

Tnd means "Industrial," C/I means "Commercial/Institutional." ET means "Engine Testing," and SWD-G means "Solid Waste Disposal-Government."

dAn "E" means the control measure is currently associated with the SCC in the CMDB, and no changes are recommended. A "D" means the control measure is
currently associated with the SCC. but this control measure should be deleted because it is not appropriate for the SCC. An "N" means the control measure is
not currently associated with the SCC in the CMDB. but adding it is recommended.

eThe control measure is assumed to be representative for the SCC: control cost data are unavailable for the specific fuel type for the SCC.

The fuel type is unknown. For the purposes of this analysis it is assumed to be a liquid because most of the emissions identified for the engine testing SCCs in
the analysis done in the Ozone T ransport Assessment Group Region were from liquid fuel-fired turbines.


-------
should be assigned. For turbines that burn miscellaneous gaseous fuels, the most representative
control measures are those for natural gas-fired turbines. Similarly, for turbines that burn
miscellaneous liquid fuels, the most representative available control measures are those for oil-
fired turbines. The description field in the CMDB table called "Table 02_Efficiencies" could be
revised to indicate that the control measures for natural gas units are assumed to be applicable for
all gaseous fuel fired units, and the control measures for oil-fired units are assumed to be
applicable for all liquid fuel-fired units (note that the separate control measures already in the
CMDB for jet fuel-fired turbines are also based on the data for oil-fired units).

Finally, gas turbine SCCs for evaporative losses from turbine fuel storage and delivery
systems are associated with NOx control measures in the current CMDB. We recommend
deleting these NOx control measure/SCC records from the CMDB table called "Table 03_SCCs"
because there should be no NOx emissions from the sources represented by these SCCs. These
control measure/SCC combinations are identified with a "D" in the applicable cells in Table 3-9.

3.4 Example Emission Limits for NonEGU Combustion Turbines

NonEGU combustion turbines are subject to sc\ era I emission regulations, including
NSPS in 40 CFR part 60 and various state regulations. Example emission limits in state
regulations are presented in Table 3-10.

Table 3-10. NOx Emissions Limits for NonEGU Combustion Turbines in New York

State

Type of Service

Type of Combustion Turbine
Operating Cycle

Emission Limit

Effective
Date

New Yorka

Any—gaseous fuel

Combined cycle

42 ppmdv (at 15% 02)

Current





Simple cycle or regenerative
cycle

50 ppmdv (at 15% 02)

Current



\n\ iiil-fiivxl

Combined cycle

65 ppmdv (at 15% 02)

Current





Simple cycle or regenerative
cycle

100 ppmdv (at 15% 02)

Current

aThe requirements apply to combustion turbines with a maximum heat input rate greater than or equal to 10 million
Btu per hour at major sources of NOx emissions. The specified limits apply until July I, 2014; beginning on July
I, 2014, owners/operators must submit a proposal for RACT (NYCRR, 2014).

3.5 References

BAAQMD, 2010. Bay Area Air Quality Management District. Preliminary Determination of
Compliance. Marsh Landing Generating Station. Application 18404. March 2010.
Available at: http://www.energy.ca.gov/sitingcases/marshlanding/documents/other/2010-
03-24 Bay Area AOMD PDOC.pdf

3-26


-------
CARB, 2004. California Environmental Protection Agency. Air Resources Board. Report to the
Legislature. Gas-Fired Power Plant NOx Emission Controls and Related Environmental
Impacts. Stationary Source Division. May 2004. Available at:
http://www.arb.ca.gov/research/apr/reports/12069.pdf

CH2MHill, 2002. Walnut Energy Center Application for Certification. Prepared for the
California Energy Commission. November 2002. Available at:

www.energy.ca.gov/sitingcases/turlock/documents/applicant files/volume 2/App 08.01
E Eval Control.pdf

Cybulski, 2006. Cybulski, A. and J. Moulin, editors. Structured Catalysts and Reactors CRC
Press. 2006. p. 236.

DeCicco, 2004. DeCicco, S., B. Reyes, and T. Girdlestone. EmeraChem, LLC. SCONOX White
Paper. Multi-Pollutant Emission Reduction Technology For Stationary Gas Turbines and
IC Engines. January 5, 2004. Available at:

\v\v\v.emera.serveyourmarket.com/papers/S('ONOx"f	tper%20-

%20rl.pdf

ECP, 2008. EmeraChem Power. Attachment in email from J. Yalmus, EmeraChem Power, to W.
Lee, BAAQMD. Request for EMx Cost Information. September 8, 2008. Available at:
http://www.baaqmd.gOv/~/media/Files/Engineering/Public%20Notices/2010/18404/Foot
notes/EMx%20BACT%20economic%20analvsis%20final09072008.ashx

EEA, 2008. Energy and Environmental Analysis (An 1CF International Company). Technology
Characterization: Gas Turbines. Prepared for Environmental Protection Agency Climate
Protection Partnership Division. December 2008. Available at:
http://www.epa.gov/chp/documents/catalog chptech gas turbines.pdf

EPA, 1993. U.S. Environmental Protection Agency. Alternative Control Techniques

Document—NOx Emissions from Stationary Gas Turbines. EPA-453/R-93-007. January
1993.

FMPA, 2004. Florida Municipal Power Agency. Chapters 3 and 4 of PSD BACT Analysis for
Stock Island Facility in Key West, Florida. Available at: Available at:
http://www.dep.state.fl.us/air/emission/construction/stockisland/BasisofBACT.pdf and
http://www.dep.state.fl.us/air/emission/construction/stockisland/NOxBACT.pdf

Kawasaki, 2010. Kawasaki Gas Turbines—Americas. Press Release. Kawasaki Gas Turbines
Cogeneration System Helps Bridgewater Correctional Facility. January 25, 2010.
Available at:

http://www.kawasakigasturbines.com/index.php/press releases/read/kawasaki gas turbi
nes cogeneration system helps bridgewater correctional fa

Leposky, 2004. Leposky, G. Oil producer installs cogeneration system with ultra-low NOx
emissions. Distributed Energy. July/August 2004. Available at:

http://www.distributedenergv.com/DE/Articles/Oil Producer Installs Cogeneration Svst
em With Ult 2857.aspx?pageid=62d52359-blc7-4115-9d0d-7837abe081cb

3-27


-------
NYCRR, 2014. New York Codes, Rules and Regulations. Title 6. Chapter III. Subchapter A.

Part 227. Subpart 227-2. Reasonably Available Control Technology (RACT) For Major

Facilities of Oxides of Nitrogen (NOx). Available at:

http:// government, westl aw. com/linkedslice/default, asp? SP=nycrr-1000

OSEC, 1999. Onsite Sycom Energy Corporation. Cost Analysis of NOx Control Alternatives for
Stationary Gas Turbines. Prepared for U.S. Department of Energy. Environmental
Programs Chicago Operations Office. November 5, 1999. Available at:
https://wwwl.eere.energy.gov/manufacturing/distributedenergy/pdfs/gas turbines nox c
ost analvsis.pdf

Peltier, 2003. Peltier, R. Gas turbine combustors drive emissions toward nil. Power Magazine.
March 15, 2003.

Quackenbush, 2012. Quackenbush, G. Cogeneration plant saves Pacific Union College $1

million a year in energy costs. North Bay Business Journal. March 19, 2012. Available at:
http://www.northbavbusinessiournal.com/50863/cogeneration-plant-saves-pacific-union-
college-1 -million-a-vear-in-energy-costs/

RDC, 2001. Resource Dynamics Corporation. Assessment of Distributed Generation Technology
Applications. Prepared for Maine Public Utilities Commission. February 2001. Available
at: http://www.distributed-generation.com/Library/Maine.pdf

RTI, 2014. Spreadsheet "Turbines control costs xlsx." Prepared for U.S. Environmental

Protection Agency Office of Air Ouulily Planning and Standards. Air Economics Group.
February 7, 2014

3-28


-------
SECTION 4
GLASS MANUFACTURING SECTOR

4.1	Introduction

The control cost database separates the glass manufacturing sector into four different
types; flat glass, container glass, pressed glass, and general glass manufacturing. The CMDB
listed six different control technologies for NOx emissions which were all reviewed in 2006 and
included cullet preheat, oxy-firing, electric boost, low NOx burners, selective catalytic reduction
(SCR) and selective non-catalytic reduction (SNCR). A literature and internet search was
conducted to find any new control technologies for NOx or any updates to existing controls
regarding cost and efficiencies. Operating permits for some glass manufacturing plants were
reviewed and control system vendors were also contacted for information. A brief summary of
data from each reference reviewed in included in the spreadsheet "CoSTGlass Mfg.xlsx."

4.2	Example NOx Regulatory Limits

4.2.1	Wisconsin

Glass manufacturing furnace with a maximum heat input capacity equal to or greater than
50 mmBtu per hour, 2.0 pounds per ton of produced glass.1

4.2.2	New Jersey

Commercial container glass, specialty container glass, borosilicate recipe glass, pressed
glass, blown glass, and fiberglass manufacturing furnaces: 4.0 lbs/ton glass removed. Flat glass
manufacturing furnaces: 9.2 lbs/ton glass removed.

4.2.3	New York

NOx emissions are covered under NY's case-by case RACT regulations.

4.3	Recommended Additions

The following NOx controls are recommended additions for the glass manufacturing
industry that are not currently in the control cost database, and a tabular summary of the costs is
presented in Table 4-1.

¦ Electric Boost—Three entries for electric boost controls were in the CMDB for

container, flat, and pressed glass manufacturing. A cost estimate for electric boost was
found for "general" glass manufacturing (DOE, 2002), since the CMDB did not have a
"general" entry for electric boost controls, an entry for "Electric Boost; Glass
Manufacturing—General" was added with a new abbreviation of NELBOGMGN. The

1 http://dnr.wi.gov/About/NRB/2007/Januarv/01-Q7-3A4.pdf

4-1


-------
reference provided an annualized cost of $7,100 per ton of NOx removed based on a
250 ton of glass per day glass melting furnace operating with an emission rate of 8-lb
NOx per ton glass produced and a NOx removal efficiency of 30 percent. Since the
reference did not provide capital costs, the capital to annual cost ratio could not be
determined, and the capital recovery factor was assumed to be the same as the electric
boost entries for container, flat, and pressed glass (i.e., 0.1424, assuming equipment
life of 10 years).

¦	Oxy-firing—Three entries for oxy-firing were in the CMDB for container, flat, and
pressed glass manufacturing. Similar to electric boost controls, an updated cost
estimate for oxy-firing was found for "general" glass manufacturing (DOE, 2002);
since the CMDB did not have a "general" entry for oxy-firing, an entry for "OXY-
Firing; Glass Manufacturing—General" was added to the CMDB with a new
abbreviation of NOXYFGMGN. The reference provided an annualized cost of $2,352
per ton of NOx removed based on a 250 ton of glass per day glass melting furnace
emitting 8-lb NOx per ton glass produced and a NOx removal efficiency of 85
percent.1 Since the capital costs were not provided the capital to annual cost ratio and
the capital recovery factor were assumed to be the same as the oxy-firing entries for
container, flat, and pressed glass, which all had the same values.

¦	Catalytic Ceramic Filter—This new control technology for NOx reduction was not
previously in the database and was added for flat glass manufacturing with a new
abbreviation of CATCFGMFT. A vendor was contacted for information (2013
Vendor Quote). The minimum and maximum cost per ton estimates were based on
regenerative gas-fired furnace with pull rates of 600 tons per day and 490 tons per
day, respectively. The estimate provided by the vendor included capital cost,
annualized capital costs, and annual operational cost in 2013 dollars; it also included
NOx reductions based on a 95 percent NOx efficiency.

Table 4-1. Summary of Cost Effectiveness and Supporting Data for Recommended
Additions

Technology

Furnace Production
Rate (ton/dav)

Cost
Year

NOx Removal
Efficiency (%)

Cost Effectiveness,
$/ton NOx

Capital to Annual
Cost Ratio

Electric Boost

250

2002

30

7,100

N/Aa

(general)











Oxy-firing

250

2002

85

2,352

2.7

General











Catalytic

490

2013

95

1,045

4.6

Ceramic Filter

600

2013

95

997

4.6

aThe ratio cannot be calculated because capital costs are not available.

4-2


-------
4.4 Recommended Changes

Changes to the CMDB are recommended for the following three types of control
measures, which are also summarized in Table 4-2.

¦	Low NOx Burner General—Three entries for Low NOX burners were in the cost
database for container, flat, and pressed glass manufacturing. An updated cost
estimate for low NOx burners for flat glass and container glass manufacturing (EC,
2013) was found and entries NLNBUGMCN and NLNBUGMFT were updated. The
reference provided capital costs and an annualized cost in euros per kilogram of NOx
removed which was converted to dollars per ton of NOx. For flat glass the minimum
cost per ton estimate was based on a 900 ton per day gas fired furnace, and the
maximum cost per ton estimate was based on a 500 ton per day gas fired furnace. For
container glass the minimum cost per ton estimate was based on 450 ton per day gas
fired furnace, and the maximum cost per ton estimate was based on a 200 ton per day
gas fired furnace. The capital recovery factor and the capital to annual cost ratio were
also updated. We also recommend changing the equipment life for low NOx burners
on flat glass furnaces from 3 years to 10 years (EC, 2013).

Additionally, equations for low NOx burners were added for entries NLNBUGMCN
and NLNBUGMFT to "Table 04_Equations" of the CMDB based on the best fit trend
lines of the total capital investment and total annual cost for the facilities with the
production levels described above, the best fit trend line results were as follows:

NLNBUGMCN (The correlation coefficients are high because the data are from a
single source, and they may reflect data points from a correlation performed by that
source)

Total capital investment (2007 dollars) = 30,930 x (tons/day)0 45 (R2 = 0.99)

Total annual cost (2007 dollars) = 9,377 x (tons/day) °'40 (R2 = 0.99)

M.NBUGMFT (The correlation coefficients are a perfect 1.0 because only two data
points are available)

Total capital investment (2007 dollars) = 527 x (tons/day) + 664,557 (R2 = 1.0)

Total annual cost (2007 dollars) = 132x (tons/day) + 150,105 (R2 = 1.0)

¦	Cullet Preheating—Two entries for cullet preheating controls were in the cost
database for container and pressed glass manufacturing. An updated annualized cost
per ton value and NOx efficiency for pressed and container glass entries (IT, 2002)
were found and updated for entries NCLPTGMCN and NCUPHGMPD. The
reference provided an annualized cost of $5,000 per ton of NOx removed based on a
250 ton of glass per day glass melting furnace emitting 8-lb NOx per ton glass
produced and a NOx removal efficiency of 5 percent.1 Since the reference did not

1 Annualized cost includes capital and O&M costs and is based on 2002 dollars.

4-3


-------
provide capital costs separately, the capital to annual cost ratio and the capital
recovery factor were not updated. Additionally, based on information from EPA's
OECA staff, class entries for cullet preheating should be changed from "known" to
"emerging" in Table OlSummary of the CMDB because this control measure is
technically feasible but has rarely been implemented.1

¦ Selective Catalytic Reduction—Three entries for selective catalytic reduction were in
the CMDB for container, flat, and pressed glass manufacturing. An updated cost
estimate for SCR for flat glass and container glass manufacturing (EC, 2013) was
found, and entries NSCRGMCN and NSCRGMFT were updated. The reference
provided capital costs and an annualized cost in euros per kilogram of NOx removed
which was converted to dollars per ton of NOx.2 For flat glass the minimum and
maximum cost per ton estimates were based on a 900 and 500 ton per day gas fired
furnace, respectively. For container glass the minimum and maximum estimates were
based on a 450 and 200 ton per day gas fired furnaces, respectively. The capital to
annual cost ratio were also updated.

Equations for SCR were added for entries NSCRGMCN and NSCRGMFT to Table 4
of the CMDB based on the best fit trend lines of the total capital investment and total
annual cost for the facilities with the production levels described above, the best fit
trend line results were as follows:

NSCRGMCN (The correlation coefficients are high because the data are from a
single source, and they may reflect data points from a correlation performed by that
source)

Total capital investment (2007 dollars) = 79,415 x (tons/day) 0 51 (R2 = 0.99)

Total annual cost (2007 dollars) = 643 x (tons/day) + 135,302 (R2 = 1.0)

NSCRGMFT (The correlation coefficients are a perfect 1.0 because only two data
points are available)

Total capital investment (2007 dollars) = 3681 x (tons/day) + 1.0E+06 (R2 = 1.0)
Total annual cost (2007 dollars) = 842 x (tons/day) + 424,930 (R2 = 1.0)

1	Personal communication. Katie McClintock, US EPA/OCEA, with Larry Sorrels, US EPA/OAR/OAQPS, Feb.
13,2014.

2	Conversion based on 2008 average exchange rate of 0.711. Source: http://www.irs.gov/Individuals/International-

Taxpavers/Yearlv-Average-Currencv-Exchange-Rates

4-4


-------
Table 4-2. Summary of Cost Effectiveness and Supporting Data for Recommended
Additions

Technology

Furnace Production
Rate (ton/day)

Cost
Year

NOx Removed
(tons/year)

Cost Effectiveness,
$/ton NOx

Capital to Annual
Cost Ratio

NLNBUGMCN

200

2007

66

1,365

4.2



450

2007

100

1,072

4.3

NLNBUGMFT

500

2007

371

574

4 :



900

2007

611

447

4.3

NCLPTGMCN

250

2002

5%

5,000

4.5

NCUPHGMPD

250

2002

5%

5,000

4.5

NSCRGMCN

200

2007

i:i

h.'J

4.5



450

2007

25 1

l.(.S4

4.2

NSCRGMFT

500

2007

SS(.



3.4



900

2007

1. W

S55

3.7

4.5	Recommended Deletions

¦	Selective Non-Catalytic Reduction Three entries for selective non-catalytic
reduction were in the cost database lor container, flat, and pressed glass
manufacturing. Based on conversations between EPA and OECA staff, SNCR entries
for glass manufacturing should be removed based on recent NSR settlements that
indicate SNCR is not a technically feasible control technology for the removal of
NOx.1

4.6	Updates to Source Classification Codes

¦	There are twenty applicable SCCs for glass manufacturing as shown in Table 4-3.

¦	In an analysis of NOx emissions for the Ozone Transport Assessment Group Region
in 2011, fourteen of the SCCs in Table 4-3 were identified. The six SCCs not
included in the Ozone Transport Region are shown at the bottom of Table 4-3. Four
of the SCCs, 30501401, 30501402, 30501403, and 30501404 are associated with
glass manufacturing NOx controls in the current CMDB.

¦	Furnaces are the primary source of NOx emissions in the glass manufacturing
industry, therefore NOx emission control techniques are typically for point emission
sources associated with furnace emissions. The four SCCs identified in the CMDB
pertain to four types of melting furnaces; general, flat, container, and pressed. The

1 Personal communication. Katie McClintock, US EPA/OCEA, with Larry Sorrels, US EPA/OAR/OAQPS, Feb. 13,
2014.

4-5


-------
remaining sixteen SCCs in Table 4-3 are not associated with furnaces. Therefore, no
changes related to SCCs are recommended for the CMDB.

¦ For new control techniques added to the CMDB for glass manufacturing, the
applicable SCC from Table 4-3 was added to the Description field for each control
technique in Table OlSummary in the CMDB (Table C-lof Appendix C of this
report). These related control measures and SCCs should also be added to "Table
03 SCCs" in the CMDB.

Table 4-3. Applicable SCCs for the Glass Manufacturing Industry

SCC Code

SCC Level One

SCC Level Two

SCC Level Three

SCC Level Four

305014013

Industrial Processes

Mineral Products

Glass Manufacture

Furnace/Genera 1 * *

305014023

Industrial Processes

Mineral Products

Glass Manufacture

Container Glass: Melting Furnace

305014033

Industrial Processes

Mineral Products

Glass Manufacture

Flat Glass: Melting Furnace

305014043

Industrial Processes

Mineral Products

Glass Manufacture

Pressed and Blown Glass:
Melting Furnace

30501406

Industrial Processes

Mineral Products

Glass Manufacture

Container Glass:
Forming/Finishing

30501407

Industrial Processes

Mineral Products

Glass Manufacture

Flat Glass: Forming/Finishing

30501408

Industrial Processes

Mineral Products

Glass Manufacture

Pressed and Blown Glass:
Forming/Finishing

30501410

Industrial Processes

Mineral Products

Glass Manufacture

Raw Material Handling (All
Types of Glass)

30501411

Industrial Processes

Mineral Products

Glass Manufacture

General **

30501413

Industrial Processes

Mineral Products

Glass Manufacture

Cullet: Crushing/Grinding

30501414

Industrial Processes

Mineral Products

Glass Manufacture

Ground Cullet Beading Furnace

30501416

Industrial Processes

Mineral Products

Glass Manufacture

Glass Manufacturing

30501420

Industrial Processes

Mineral Products

Glass Manufacture

Mirror Plating: General

30501499

Industrial Processes

Mineral Products

Glass Manufacture

See Comment **

SCCs Not Included in the Ozone Transport Assessment Group Region:



30501405

Industrial Processes

Mineral Products

Glass Manufacture

Presintering

30501412

Industrial Processes

Mineral Products

Glass Manufacture

Hold Tanks **

30501415

Industrial Processes

Mineral Products

Glass Manufacture

Glass Etching with Hydrofluoric
Acid Solution

30501417

Industrial Processes

Mineral Products

Glass Manufacture

Briquetting

30501418

Industrial Processes

Mineral Products

Glass Manufacture

Pelletizing

30501421

Industrial Processes

Mineral Products

Glass Manufacture

Demineralizer: General

aDenotes SCCs included in the CMDB.

4-6


-------
4.7 References

DOE, 2002. Oxygen Enriched Air Staging a Cost-effective Method For Reducing NOx

Emissions. U.S. Department of Energy. Office of Industrial Technologies. April 2002.

http://wwwl .eere.energy, gov/manufacturing/resources/

glass/pdfs/airstaging.pdf

EC, 2013. Best Available Techniques (BAT) Reference Document for the Manufacture of Glass.
European Commission 2013. http://eippcb.irc.ec.europa.eu/reference/BREF/
GLS Adopted 03 2012.pdf

Vendor Quote 2013 - Confidential Business Information

4-7


-------
SECTION 5
LEAN BURN ENGINES

The CMDB includes the following NOx emissions control measures for Lean Burn
Engines:

¦	Air to fuel ratio (AFR) (achieves 20 percent reduction)

¦	Air to fuel ratio (AFR) and Ignition retard (IR) (achieves 30 percent reduction)

¦	Ignition retard (IR) (achieves 20 percent reduction)

¦	Low emission combustion (achieves 87 percent reduction)

¦	Low emissions combustion, low speed (achieves 87 percent reduction)

¦	Low emissions combustion, medium speed (achieves 87 percent reduction)

¦	Nonselective catalytic reduction (NSCR) (achieves 90 percent reduction)

¦	Selective catalytic reduction (SCR) (achic\ cs on percent reduction)

¦	Selective noncatalytic reduction (SNCR) (achieves 90 percent reduction)

Based on the literature review and the new cost data identified for Lean Burn control
technologies, several changes to the CMDB are recommended. No changes to existing records in
CMDB are recommended. The following sections outline the additions and other comments
recommended for the CMDB in relation to NOx emissions from Lean Burn Engines.

5.1 Literature Search

In order to update the existing control measures database, a literature search was
conducted for articles and papers published since 2008 (to include 2008 through August 2013)
using the following terms:

¦	engine

¦	lean burn

¦	cost

¦	NOx or "nitrogen oxides"

¦	scr or "selective catalytic reduction"

¦	turbocharge

5-1


-------
¦	air/fuel ratio

¦	layered combustion

¦	high energy ignition

¦	high pressure fuel injection

¦	"low emission control" or LEC

¦	electronic engine control

¦	combustion modification

¦	timing

¦	exhaust gas recirculation

¦	lean NOx catalyst

¦	lean NOx trap

¦	control efficiency

¦	emission reduction

The literature search identified a total of 19 references, and the abstracts for these
references were reviewed. Three references of potential interest were identified and two of these
were obtained for review in the lean burn engine control device study.

5.2 Document Review

A brief summary of data from each reference reviewed in included in the spreadsheet
"CoSTleanburn.xlsx," in worksheet "Overall Sum—New Ref Review." The information and
data available from each reference is provided in table format, along with indication of whether
the data were used or not.

There are 6 control technique additions to be added to the CMDB from 5 references.

The recommended additions include:

¦	Low Emission Combustion, LEC (for natural gas engines);

¦	Layered Combustion, LC (for 2 stroke natural gas engines);

¦	Layered Combustion, LC (for 2 stroke Large Bore natural gas engines);

5-2


-------
¦	Air to Fuel Ratio Controller, AFRC;

¦	Selective Catalytic Reduction, SCR (for 4 stroke natural gas engines); and

¦	SCR (for diesel engines).

Recent cost data for these control techniques were available from reports dated 2001
through 2012.

The references for the added control techniques are included on the "Table 06
References" worksheet and are as follows:

OTC 2012. Technical Information Oil and Gas Sector, Significant Stationary Sources of NOx
Emissions. Final. October 17, 2012.

SJVAPCD 2003. RULE 4702—Internal Combustion Engines—Phase 2. Appendix B, Cost
Effectiveness Analysis for Rule 4702 (Internal Combustion Engines—Phase 2). San
Joaquin Valley Air Pollution Control District. July 17, 2003.
www, arb. ca. gov/pm/pmmeasures/ceffect/rul es/si vaped 4702. pdf

CARB 2001. Determination of Reasonably Available Control Technology and Best Available
Retrofit Control Technology for Stationary Spark-Ignited Internal Combustion Engines.
California Environmental Protection Agency, Air Resources Board, Stationary Source
Division, Emissions Assessment Branch, Process Evaluation Section. November 2001.

EPA 2010. Alternative Control Techniques Document: Stationary Diesel Engines. March 5,
2010.

PA DEP 2013. Technical Support Document General Permit GP-5. Pennsylvania Department of
Environmental Protection. Bureau of Air Quality. January 31, 2013.

5.3 Low Emission Combustion (LEC) (NLECICENG)

The costs and cost effectiveness for applying LEC to natural gas Lean Burn engines are
obtained from the document Appendix B, Cost Effectiveness Analysis for Rule 4702 (Internal
Combustion Engines Phase 2) (SJVAPCD 2003). Information was provided on Capital costs,
Annual costs, uncontrolled emissions, and reduction efficiency. The assumptions for the original
reference analysis are provided in Table 5-1 for LEC along with changes in assumptions for the
current analysis.

LEC are described as retrofit kits that allow engines to operate on extremely lean fuel
mixtures to minimize NOx emissions. The LEC retrofit may include: (1) redesign of cylinder
head and pistons to improve mixing (on smaller engines), (2) Precombustion chamber (on larger
engines), lower cost, simple versions, (3) Turbocharger, (4) High energy ignition system,

5-3


-------
Table 5-1. LEC for Natural Gas Lean Burn Engines

Assumptions in Original Reference

Changes to Assumptions Made in Current Analysis

Capital costs

Control efficiency: 80%

Capital costs: provided for multiple models
Annual Costs

Equipment life: 10 yr
Interest rate: 10%

Operating hours: 2000 hr/yr
Emission rate, uncontrolled: 740 ppmv

Emission rate, controlled: 80% reduction

Annualized equipment cost: provided for multiple
model sizes

None.

None.

None.

Interest rate: 7%

None.

Emission rate, uncontrolled: Assumed in id 10 upper end
hp rating for each model.

Xone

( Kl u 1424

Annual O&M cost: assumed $0.

None

(5) Aftercooler, and (6) Air to fuel ratio controller. (A discussion of individual technologies is
provided in Appendix B of the original reference, pp. B-l to B-28). No detail was provided on
the exact combination of combustion modifications included in the example cost analysis; some
references indicate that LEC on larger engines often includes aPCC (p.B-10) (CARB 2001).
LEC are known or demonstrated control techniques for lean burn engines. An 80 percent NOx
emission reduction can be achieved by LEC with little or no fuel penalty (in fact, LEC
technologies are expected to decrease fuel consumption because they result in leaner burning
engine, though the costs do not account for fuel consumption decrease). The original reference
assumed an 80 percent reduction in the cost example.

Capital and annual costs were provided for multiple size ranges of engines. The capital
costs ranged from $14,000 to $256,000. Costs for the 1000 to 3000 hp model were given as
$40,000 to $256,000, and a mid-range cost of $148,000 was assumed in the current analysis. The
total annual costs ranged from $2,000 to $21,000 (these costs are very similar to the costs
calculated in the original reference analysis). The original reference assumed there are no annual
operation and maintenance costs incurred from the combustion modification technologies, and
the only annual cost provided is for annualized capital costs. No emission reductions are
provided in the document (however the final cost effectiveness values are provided and the
reduction assumed in the original analysis can be back-calculated). In the current analysis, a hp

5-4


-------
rating based on the middle or upper end of each size range was assumed for estimating the
uncontrolled NOx emissions. An estimate of emissions was made in the current analysis.
Uncontrolled NOx emissions were estimated based on an uncontrolled NOx concentration of 740
ppmv (this equates to approximately 9 g/bhp-hr), the operating hours were provided as 2000
hr/yr in the original reference, and controlled emissions were estimated based on 80 percent
reduction as stated in the reference. Uncontrolled NOx emissions ranged from 1.1 to 34 tpy for
the models, and the NOx reductions ranged from 0.90 to 27 tpy for the models

The current analysis shows a cost effectiveness of $2,200/ton of NOx reduction lo
$780/ton for 2000 hr/yr operation, and the average cost effectiveness across all the models is
$l,000/ton of NOx reduction.

The cost year is not provided in the reference; assumed the cost year is the date of the
cited reference, 2001$.

Based on the cost calculations for engines of varying hp, the following equations were
developed for the capital cost and annual costs for l.F.C on natural gas Lean Burn engines:

Capital cost = 16019 e 0 00,6x(hP)

Annual cost = 2280.8 e 0 0016x(hP)

The R2 value for these equations is 0.96. These equations should be included in the CoST
database file under a new equation type.

See the cost calculations in worksheet "LEC (CARB)-2001" of the Excel file.
5.4 Layered Combustion (LC), 2 Stroke (NLCICE2SNG)

The costs and cost effectiveness for applying LC to natural gas Lean Burn engines (2
stroke) are obtained from the document Technical Information Oil and Gas Sector, Significant
Stationary Sources of NOx Emissions (OTC 2012). Information was provided on Capital costs;
assumptions were made to determine Annual costs, uncontrolled emissions, and reduction
efficiency. The assumptions for the original reference analysis are provided in Table 5-2 for LC
for 2 stroke engines, along with changes in assumptions for the current analysis.

LC consists of multiple combustion modification technologies. The combustion
modifications included in this example are related to (1) Air supply; (2) Fuel supply, (3) Ignition,
(4) Electronic controls, and (5) Engine monitoring (a discussion of individual technologies is
provided on pp. 17 to 19 for 2 stroke Lean Burn engines). No significant detail was provided on
which specific combustion modification technologies were applied. In the example study, 3 of

5-5


-------
the most representative manufacturer and models of 2 stroke Lean Burn engines used for integral
compressors were selected for evaluation; these 3 engines were Cooper GMVH-10 2250 hp,

Table 5-2. LC for Natural Gas Lean Burn Engines, 2-stroke

Assumptions in Original Reference
Capital costs

Control efficiency: Not provided.

Capital costs: based on cited ERLE 2009 project
(First unit upgrade costs)

Annual Costs

Equipment life: Not provided.

Interest rate: Not provided.

Operating hours: Not provided.

Emission rate, uncontrolled: Not provided
Emission rate, controlled: 0.5 g/bhp-hr
Annualized equipment cost: Not provided.
Annual O&M cost: Not provided.

Changes to Assumptions Made in Current Analysis

Control efficiency: derived value is 97% (this is high)

Capital costs: used average based on the provided range
for each make/model engine.

Equipment life: 10 yr

Interest rale: 7%

Operating hours: 2000 lvr/yr

Emission rale, uncontrolled: 16.8 g/bhp-hr

None.

CRF: 0.1424
Annual O&M cost: $0.

Clark TLA-6 2000 hp, and Cooper GMW-10 2500 hp (cited ERLE 2009 report "ERLE Cost
Study of the Retrofit Legacy Pipeline Engines to Satisfy 0.5 g/bhp-hr NOx"). LC are known or
demonstrated control techniques for lean burn, 2 stroke engines. A NOx emissions rate of 0.5
g/bhp-hr was achieved. The OTC 2012 document provided an estimate of the capital cost range
for retrofitting technologies to achie\ c the outlet NOx limit for each engine. An average cost
based on the range was estimated lor each engine and used in the current analysis. Details on the
buildup of these costs are not provided in the OTC 2012 document. No annual costs are provided
in the document. No emission reductions are provided in the document.

Based on the review of other references in this analysis, it was assumed that there are no
additional annual operating costs incurred from the combustion modification technologies,
except for annualized capital costs (CARB 2001). Because no emission reduction data were
provided, an estimate of emissions was made in the current analysis. Uncontrolled NOx
emissions were assumed to be 16.8 g/bhp-hr (EPA 2003), controlled emissions were 0.5 g/bhp-hr
as stated in the reference, and the operating hours were assumed to be 2000 hr/yr (this
assumption is consistent with the LEC operating hours in the CARB 2001 document).

5-6


-------
Uncontrolled NOx emissions for the 3 similar sized engines ranged from 74 to 93 tpy, and the
NOx reductions ranged from 72 to 90 tpy.

Based on the 3 make and model engines, the average cost was estimated to be $2,800,000
for approximately 2250 hp engines (average hp of the 3 units), and the average total annual cost
was estimated to be $390,000. The average cost effectiveness is $4,900/ton of NOx reduction for
2000 hr/yr operation.

The cost year is not provided in the reference; we assumed the cost year is ihe dak- of the
cited Cameron 2010 retrofit project, 2010$.

See the cost calculations in worksheet "Overall Sum—New Ref Review" of the Excel
file, rows 21 through 25.

5.5 Layered Combustion (LC), Large Bore, 2 Stroke, Low Speed (NLCICE2SLBNG)

The costs and cost effectiveness for applying LC to natural gas Lean Burn engines (2
stroke Large Bore) are obtained from the document Technical Information Oil and Gas Sector,
Significant Stationary Sources of NOx Emissions (OK' 2<) 12). Large Bore RICE are those with
large piston diameters. The larger the bore (or piston diameter), the larger the volume available
for engine combustion, and hence the greater the power delivered by the engine. Information was
provided on Capital costs; assumptions were made to determine Annual costs, uncontrolled
emissions, and reduction efficiency. The assumptions for the original reference analysis are
provided in Table 5-3 for 1 ,C for large bore 2 Stroke engines, along with changes in assumptions
for the current analysis

LC consists of multiple combustion modification technologies. The combustion
modifications included (1) High pressure fuel injection; (2) Turbocharging, (3) Precombustion
chamber, and (4) Cylinder head modifications (a discussion of individual technologies is
provided on pp. 18 to 19 for 2 stroke Lean Burn engines). LC are known or demonstrated control
techniques for lean burn, large bore, 2 stroke engines. These modifications achieved a NOx
emissions rate of 0.5 g/bhp-hr. The OTC 2012 document provided ranges of capital costs for
retrofitting combustion modifications for large bore 2 stroke Lean Burn engines from 200 to
11,000 hp (cited Cameron 2011 presentation "Available Emission Reduction Technology for
Existing Large Bore Slow Speed Two Stroke Engines." A copy of this presentation was not
found.). Details on the buildup of these costs are not provided in the OTC 2012 document. No
annual costs are provided in the document. No emission reductions are provided in the
document.

5-7


-------
Based on the review of other references in this analysis, it was assumed that there are no
additional annual operating costs incurred from the combustion modification technologies,
except for annualized capital costs (CARB 2001). Because no emission reduction data were
provided, an estimate of emissions was made in the current analysis. Uncontrolled NOx

Table 5-3. LC for Natural Gas Lean Burn Engines, Large Bore 2-stroke

Assumptions in Original Reference

Changes to Assumptions Made in Current Analysis

Capital costs

Control efficiency: Not provided.

Capital costs: based on cited Cameron 2010
project

Annual Costs

Equipment life: Not provided.

Interest rate: Not provided.

Operating hours: Not provided.

Emission rate, uncontrolled: Not provided
Emission rate, controlled: 0.5 g/bhp-hr
Annualized equipment cost: Not provided.
Annual O&M cost: Not provided.

Control efficiency: derived value is 97% (this is high)
None.

Equipment life: 10 yr

Interest rale: 7%

Operating hours: 2.000 hr/yr

Emission rale, uncontrolled: 16.8 g/bhp-hr

None

( Kl u 1424
Annual O&M cost: $0.

emissions were assumed to be 16.8 g/bhp-hr (EPA 2003), controlled emissions were 0.5 g/bhp-hr
as stated in the reference, and the operating hours were assumed to be 2000 hr/yr (this
assumption is consistent with the LEC operating hours in the CARB 2001 document).
Uncontrolled NOx emissions were estimated to be 410 tpy for the larger 11,000 hp engines and
were estimated to be 7.4 tpy for the smaller 200 hp engines.

For the larger 11,000 hp engines, the current analysis shows a cost effectiveness of
$l,500/ton of NOx reduction, and for the smaller 200 hp engines, the cost effectiveness is
$38,000/ton of NOx reduction.

The cost year is not provided in the reference; assumed the cost year is the date of the
cited Cameron 2010 retrofit project, 2010$.

See the cost calculations in worksheet "Overall Sum—New Ref Review" of the Excel
file, rows 12 and 13.

5-8


-------
5.6 Air to Fuel Ratio Controller (AFRC) (NAFRCICENG)

The costs and cost effectiveness for applying AFRC to natural gas Lean Burn engines are
obtained from the document Determination of Reasonably Available Control Technology and
Best Available Retrofit Control Technology for Stationary Spark-Ignited Internal Combustion
Engines (CARB 2001). Information was provided on Capital costs; assumptions were made to
determine Annual costs, uncontrolled emissions, and reduction efficiency. The assumptions for
the original reference analysis are provided in Table 5-4 for AFRC, along with changes in
assumptions for the current analysis.

Table 5-4. AFRC for Natural Gas Lean Burn Engines

Assumptions in Original Reference
Capital costs

Control efficiency: not provided
Capital costs: provided for multiple models
Annual Costs

Equipment life: 10 yr
Interest rate: 10%

Operating hours: 2000 hr/yr
Emission rate, uncontrolled: 740 ppmv

Emission rate, controlled: 80% reduction

Annualized equipment cost pan ided for multiple
model sizes

Annual O&M cost: assumed V).

Changes to Assumptions Made in Current Analysis

Control efficiency: assumed 20%

None.

None.

Interest rate: 7%

None.

Emission rate, uncontrolled: Assumed mid to upper end
hp rating for each model.

None.

CRF: 0.1424
None.

AFRC are electronic engine controls that typically monitor engine parameters and
atmospheric conditions to determine the correct air/fuel mixture for the operating condition, such
as varying engine load or speed conditions, varying ambient conditions, or startup/shutdown
conditions. (OTC 2012) (A discussion of individual technologies is provided in Appendix B of
the original reference, CARB 2001, pp. B-l to B-28). AFRC are known or demonstrated control
techniques for lean burn engines. An 80 percent NOx emission reduction can be achieved by
AFRC in combination with other combustion modifications, however a fuel consumption penalty
of up to 3 percent can occur due to AFRC.

Capital were provided for multiple size ranges of engines. The capital costs ranged from
$4,200 to $6,500 per engine.

5-9


-------
No annual costs were provided in the document. No emissions reductions were provided
in the document. Based on the cost analysis for other combustion technology controls in this
document, it was assumed that there are no additional annual operating costs incurred from the
combustion modification technologies, except for annualized capital costs (this assumption
ignores the fuel penalty issue). The total annual costs ranged from $600 to $930. Because no
emission reductions were provided in the document, an estimate of emissions was made in the
current analysis. In the current analysis, a hp rating based on the middle or upper end of each size
range was assumed for estimating the uncontrolled NOx emissions. Uncontrolled NOx emissions
were estimated based on an uncontrolled NOx concentration of 740 ppmv (this equates to
approximately 9 g/bhp-hr), the operating hours were assumed to be 2000 hr/yr (similar to the
operating hours for other control technology analyses provided in the document), and controlled
emissions were estimated based on an assumption of 20 percent reduction. Uncontrolled NOx
emissions ranged from 1.1 to 34 tpy for the models, and the NOx reductions ranged from 0.22 to
6.7 tpy for the models.

The current analysis shows a cost effectiveness of $2,700/ton of NOx reduction to
$140/ton for 2000 hr/yr operation, and the average cost effectiveness across all the models is
$810/ton of NOx reduction.

The cost year is not provided in the reference; assumed the cost year is the date of the
cited reference, 2001$.

Based on the cost calculations for engines of varying hp, the following equations were
developed for the capital cost and annual costs for AFRC on natural gas Lean Burn engines:

Capital cost= 1.3007 x (hp) + 4354.5

Annual cost = 0.1852 x (hp) + 619.99

The R2 value for these equations is 0.87. These equations should be included in the CoST
database file under a new equation type.

See the cost calculations in worksheet "AFRC (CARB)-2001" of the Excel file.
5.7 SCR (for 4 Stroke Natural Gas Engines) (NSCRICE4SNG)

The costs and cost effectiveness for applying SCR to natural gas engines are obtained
from the document Appendix B, Cost Effectiveness Analysis for Rule 4702 (Internal Combustion

5-10


-------
Engines—Phase 2) (SJVAPCD 2003). Information was provided on Capital costs, Annual costs,
uncontrolled emissions, and reduction efficiency. The assumptions for the original reference
analysis are provided in Table 5-5 for SCR for natural gas engines along with changes in
assumptions for the current analysis. SCR is a known or demonstrated control technique for lean
burn engines, although multiple references indicate that the feasibility of SCR application for
lean burn engines is highly site-specific.

Table 5-5. SCR for Natural Gas Lean Burn Engines, 4-stroke.

Assumptions in Original Reference
Capital costs

Control efficiency: 90%

Capital costs: based on RACT/BARCT
Determination.

Annual Costs

Equipment life: 10 years
Interest rate: 10%

Operating hours rate 1: 2190 hr/yr (equivalent to
capacity factor of 0.25)

Operating hours rate 2: 6570 hr/vr (cquivalciii lo
capacity factor of 0.75)

Emission rate, unconlrollcd: 740 ppmv NOx

Emission rate, controlled: 65 ppmv NOx

Annualized equipment cost: based on
RACT/BARCT Determination.

Annual O&M cost: based on RACT/BARCT
Determination.

Changes to Assumptions Made in Current Analysis

None.

None

None

lnkTcsl rale
None

\i»ne

None.

None.

None.

None.

The installed equipment capital cost ranged from $45,000 to $185,000 for 50 hp engines
and 1500 hp engines, respectively. The total annual costs ranged from $27,000 for a 50 hp
engine to $ 140,000 for a 1500 hp engine (these costs are very similar to the costs calculated in
the original reference analysis; the only difference in annual costs is related to the CRF, i.e.,
changing the interest rate from 10 percent in the original reference analysis to 7 percent in the
current analysis). NOx emissions are provided for two cases: a capacity factor of 0.25 (2190
hr/yr) and a capacity factor of 0.75 (6570 hr/yr). The uncontrolled NOx emissions ranged from
1.2 to 37 tpy for the lower capacity case, and the NOx reductions ranged from 1.1 to 33 tpy. For
the higher capacity case, uncontrolled NOx emissions ranged from 3.7 to 110 tpy, and the NOx
reductions achieved ranged from 3.3 to 100 tpy. The current analysis shows an average cost

5-11


-------
effectiveness of $8,700/ton of NOx reduction for 2190 hr/yr of operation, and $2,900/ton of NOx
reduction for 6570 hr/yr operation (these cost effectiveness values are very similar to the costs
shown in the original reference analysis).

Based on the cost calculations for engines of varying hp and annual capacity operating,
the following linear equations were developed for the capital cost and annual costs for SCR on
natural gas 4-stroke lean burn engines:

Capital cost = 107.1 x (hp) + 27186

Annual cost = 83.64 x (hp) + 14718

The R2 values for these equations are 0.95 for capital cost and 0.98 for annual cost. These
equations should be included in the CoST database file under a new equation type for linear
equations.

The cost year is not provided in the reference: assumed the cost year is the date of the
cost-basis document, 2001$.

See the cost calculations in worksheet "SCR NG (SJVAPCD)-2003" of the Excel file.
[Other cost effectiveness values for SCR are available from the PA DEP that are higher than the
cost effectiveness values shown for the SJ VAPCD SCR analysis, and other analyses. See the
summary of SCR costs in worksheet "Other SCR Cost Info" of the Excel file.]

5.8 SCR (for Diesel Engines) (NSCRICEDS)

The costs and cost effectiveness for applying SCR to diesel lean burn engines is provided
in Alternative Control Techniques Document: Stationary Diesel Engines (EPA 2010). The
assumptions for the original reference analysis are provided in Table 5-6 for SCR for diesel
engines, along with changes in assumptions for the current analysis. SCR is a known or
demonstrated control technique for lean burn, diesel engines.

Approximately 76 percent of the population of stationary diesel engines is less than 300
hp and the remaining 24 percent is greater than 300 hp. Applications for stationary engines under
300 hp include standby power generation, agriculture, and industrial applications, and less than 5
percent are used for continuous power generation. Applications for stationary engines greater
than 300 hp are primarily power generation and are almost evenly divided between continuous
duty and standby applications.

5-12


-------
The cost analysis provided in the original reference includes an assumption that
stationary diesel lean burn engines operate approximately 1000 hr/yr. This assumption is likely
appropriate for the majority of those units that are less than 300 hp and for half of the diesel
engines greater than 300 hp, i.e., approximately 87 percent of diesel lean burn engines (this
ignores the "fewer than 5 percent" used for continuous power generation). For the remaining 13
percent of engines that are greater than 300 hp and used in continuous power generation
applications, an assumption for longer operating hours, such as 8000 hr/yr, may be needed to
estimate the cost effectiveness.

Table 5-6. SCR for Diesel Lean Burn Engines—Assumptions

Assumptions in Original Reference	Changes to Assumptions Made in Current Analysis
Capital costs

Control efficiency: 90 %	None.

Equipment life: 15 year	None.

Interest rate: 7%	None

Capital costs: $98/hp	None
Annual Costs

Operating hours: 1000 hr/yr	None.

Annual costs: $40/hp (based on 1000 hr/vr	None.

operation; already includes Capital Recovery)

The original reference analysis provided a capital cost of $98/hp, and based on the mid-
range hp rating for four model engines, the capital costs ranged from $7,300 to $98,000 for SCR.
The original reference analysis provided an annual cost of $40/hp, and the annual costs ranged
from $3,000 to $40,000 per year. Uncontrolled NOx emissions factors in the original reference
were based on Tier 0 to Tier 3 values1 and an assumption of 1000 hr/yr operation. Uncontrolled
NOx emissions range from 0.25 to 9.2 tpy across the four models, and the NOx reductions
ranged from 0.22 to 8.3 tpy.

The current analysis shows an average cost effectiveness of $9,300/ton of NOx reduction
for 1000 hr/yr of operation (no weighting to the average based on engine age was applied). The
cost effectiveness over the engine size range varied from $4,800/ton to $16,000/ton for diesel
engines (and are very similar to the costs shown in the original reference analysis). It is

1 Federal Standards, from the Exhaust and Crankcase Emission Factors for Nonroad Engine Modeling—
Compression Ignition. EPA Publication No. EPA420-P-04 009. April 2004.

5-13


-------
important to note that the cost effectiveness is correlated to the manufacturing year of the diesel
engine, i.e., the Tier limit for NOx emissions. Older engines manufactured prior to 1998 have the
most lenient emissions limit while later model years have more stringent NOx emission limits
(lower baseline emissions). The overall magnitude of emission reduction achieved by the SCR is
lower for later model years as compared to earlier years, and therefore, the cost effectiveness
values are higher for later model years.

[Note: This analysis shows emission reductions and cost effectiveness lor existing and
new diesel engines through approximately 2011, the last year for phase in of the Tiered emission
values. The original reference provided information (circa 2005) on the age of the stationary
engine population, with approximately 57 percent of engines at that time being manufactured
prior to 1994 and approximately 42 percent manufactured after 1994 (note that the grouping of
the age data does not align well with the Tier years, in that the age data shows breaks in 1994
and 2003 while the Tier ranges show breaks in 1996, 1998, 2002, 2003, etc.). As the diesel
engine population continues to age and older engines are retired (i.e., those diesel engines subject
to the Pre-1998 and the Tier 1 (1998 to 2003) or Tier 1 (1996 to 2001), etc. and are replaced with
newer engines achieving lower NOx baseline emissions, the cost effectiveness for new engines
would tend to be in the higher end shown for each model and would contribute to a somewhat
higher average cost-effectiveness value. The average cost effectiveness will likely move toward
the $13,000/ton to $16,000/ton of NOx reduction range.]

See the cost calculations in worksheet "SCR Diesel (EPA Dies ACT)-2010" of the Excel

file.

5.9 Applicable SCCs for Lean Burn Engine Control Measures

Table 5-7 lists all of the ICE SCCs that are associated with one or more gas lean burn
ICR control measures in the CMDB table called "Table 03_SCCs." These SCCs were identified
with NOx emissions in an EPA query of the NEI for facilities in the Ozone Transport Group
Assessment Region (i.e., 37 states that are partially or completely to the east of lOOoW
longitude). The control measures shown in the column headings in Table 5-7 are the ICE control
measures that are currently in the CMDB. Each control measure that was determined to be
applicable for a specific SCC is identified with an "N" in the cell, meaning the control measure is
"new," i.e., not currently linked to this particular SCC, but we recommend adding this link in the
database. In some cases, we recommend applying new links between existing control measures
and existing SCCs. For example, some of the SCCs are for ICE that are fired with relatively
uncommon fuels such as process gas, methanol, digester gas, or landfill gas. While we have not

5-14


-------
Table 5-7. Potential Reciprocating Engine SCCs to Add to the CMDB and Applicable Control Measures

Applicable Conl ml Measures lur the Reciprocating Engine SCCd

SCCa

SCC

Level SCC
lb Level 2C

in

O
U

in

o
u

SCC Level 3

SCC Level 4





§
U

HH

£5
/¦

o
u

o

o

d
o

u

SS s s

§
u

HH

M

u

{*)

l/l

o
u

HH

M

u

{*)

hJ

o

u

HH

u

d
o

u

u

o
u

HH

u

z

20200702

ICE

Ind

Process Gas

Reciprocating Engine







N



N

N

20200712

ICE

Ind

Process Gas

Reciprocating: Exhaust

\

\



N



N

N

20201602

ICE

Ind

Methanol

Reciprocating Engine









N





20201607

ICE

Ind

Methanol

Reciprocating: Exhaust





\



N





20201702

ICE

Ind

Gasoline

Reciprocating Engine





N



N





20201707

ICE

Ind

Gasoline

Reciprocating: Exhaust





N



N





20280001

ICE

Ind

Equipment Leaks

Equipment Leaks















20282001

ICE

Ind

Wastewater, Aggregate

Process Area Drains















20300702

ICE

C/I

Digester Gas

Reciprocating: POTW Digester Gas

N

N



N



N

N

20300707

ICE

C/I

Digester Gas

Reciprocating: Exhaust

N

N



N



N

N

20300802

ICE

C/I

Landfill Gas

Reciprocating

N

N



N



N

N

20400401

ICE

ET

Reciprocating Engine

Gasoline





N



N





20400402

ICE

ET

Reciprocating Engine

Diesel/Kerosene





N

N

N N

N

N

20400404

ICE

ET

Reciprocating Engine

Process Gas

N

N



N





N

20400406

ICE

ET

Reciprocating Engine

Kerosene/Naphtha (Jet Fuel)





N



N





20400409

ICE

ET

Reciprocating ] ingine

Liquified Petroleum Gas (LPG)





N



N





aSCCs represent reciprocating engine activities thai were identified with NOx emissions in the Ozone Transport Assessment Group Region analysis but are not

associated with reciprocating engine control measures in the current CMDB.
bICE means "Internal Combustion Engines."

°Ind means "Industrial," C/I means "Commercial/Institutional," and ET means "Engine Testing."

dThe control measure is assumed to be representative for the SCC; control cost data are unavailable for the specific fuel type for the SCC.


-------
located any analyses that determined the applicable controls and related costs for ICE fired with
such fuels, similar control measures can be assigned to these SCCs. In order to conduct CoST
modeling analyses for these ICE, the most representative available control measures could be
assigned. For ICE that burn miscellaneous gaseous fuels, the most representative control
measures are those for natural gas-fired ICE. Similarly, for ICE that burn miscellaneous liquid
fuels such as methanol, gasoline, kerosene/diesel, and LPG, the most representative available
control measures are those for gas- or diesel-fired ICE. Also, for ICE that burn liquid fuels such
as diesel/kerosene, the most representative available control measures are those for gas-, diesel-,
or oil-fired ICE.

Six new control measures have been added to the CMDB for lean burn engines under this
review and these control measures are described in Sections 5.3 through 5.8 of this report.

Table 5-8 lists those SCCs that should be associated with the newly added lean burn engine
control measures. Each control measure that was determined to be applicable for a specific SCC
is identified by a "Y," which means yes.

In Table 5-9, a number of recommendations were made to delete NOx control
measure/SCC combinations from the CMDB. ICE SCCs for evaporative losses from fuel storage
and delivery systems are incorrectly associated with NOx control measures in the current
CMDB, and we recommend deleting these all NOx control measure/SCC records from the
CMDB table called "Table 03_SCCs" because there should be no NOx emissions from the
sources represented by these SCCs. In addition, multiple ICE control measures are misassigned
to turbine SCCs and we recommend deleting these NOx control measure/SCC records. The
reverse issue also exists where multiple turbine control measures are misassigned to ICE SCCs,
and we recommend deleting these NOx control measure/SCC records, as well. These control
measure/SCC combinations are identified in Table 5-9.

5.10 Pennsylvania General Permit 5 (GP-5) for Natural Gas Compression and/or

Processing Facilities

Pennsylvania DEP recently released a general permit for Natural Gas Compression
and/or Processing Facilities that includes limits on NOx emissions from ICE. NOx emission
limits from this general permit, along with other NOx limits for Pennsylvania, are shown in
Table 5-10. Typical emission rates and the cost-effectiveness values for applying certain control
measures are shown for lean burn and rich burn engines in Table 5-11.

5-16


-------
Table 5-8. Recommended New Control Measures to Associate With Lean Burn Reciprocating Engine SCCs in the CMDB

Applicable Control Measures for the Lean Burn
Reciprocating Engine SCC











O

z

w
y

O

z

in

r\
W

O
z

CO

hJ

l/l
r\
W

O

z

w
U

HH

u

O

z

l/l

w
y

VI

O

w

y

scca

SCC Level 1

SCC Level 2

SCC Level 3

SCC Le\ el 4

u
w
-j

z

u

HH

u

hJ

z

u

HH

u

hJ

z

g

zi

5

u

i/i

z

5

u

VI

Z

20200102

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Reciprocating











Y

20200107

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Reciprocating: Exhaust











Y

20200252

Internal Combustion Engines

Industrial

Natural Gas

2-cycle Lean Burn

Y

Y

Y

Y

Y



20200254

Internal Combustion Engines

Industrial

Natural Gas

4-cycle Lean Burn

Y

Y

Y

Y

Y



20200255

Internal Combustion Engines

Industrial

Natural Gas

2-cycle Clean Burn

Y

Y

Y

Y

Y



20200256

Internal Combustion Engines

Industrial

Natural Gas

4-cycle Clean Burn

Y

Y

Y

Y

Y



2020040lb

Internal Combustion Engines

Industrial

Large Bore Engine

Diesel





Y







20200402b

Internal Combustion Engines

Industrial

Large Bore Engine

Dual Fuel (Oil/Gas)





Y







20200403 b

Internal Combustion Engines

Industrial

Large Bore Engine

Cogeneration: Dual Fuel





Y







aSCCs represent reciprocating engine activities thai were identified with NO.\ emissions in the recent Ozone Transport Region analysis but are not associated

with reciprocating engine control measures in the current CMDB.
bThe control measure is assumed to be representative for the SCC: control cost data arc unavailable for the specific fuel type for the SCC.


-------
Table 5-9. Recommended Control Measure Deletions From SCCs in the CMDB

see

SCC Level 1

SCC Level 2

SCC Level 3

SCC Level 4

Control Measures
Recommended for Deletion

20200106

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Reciprocating:
Losses

Evap

All NOx control measures

20200206

Internal Combustion Engines

Industrial

Natural Gas

Reciprocating:
Losses

Evap

All NOx control measures

20200306

Internal Combustion Engines

Industrial

Gasoline

Reciprocating:
Losses

Evap

All NOx control measures

20200406

Internal Combustion Engines

Industrial

Large Bore Engine

Reciprocating:
Losses

Evap

All NOx control measures

20200506

Internal Combustion Engines

Industrial

Residual/Crude Oil

Reciprocating:
Losses

Evap

All NOx control measures

20200906

Internal Combustion Engines

Industrial

Kerosene/Naphtha (.lei Fuel)

Reciprocating:
Losses

Evap

All NOx control measures

20201006

Internal Combustion Engines

Industrial

Liquified Petroleum Gas (LPG)

Reciprocating:
Losses

Evap

All NOx control measures

20300106

Internal Combustion Engines

Commercial/Institutional

Distillate Oil (Diesel)

Reciprocating:
Losses

Evap

All NOx control measures

20300206

Internal Combustion Engines

Commercial/Institutional

Natural Gas

Reciprocating:
Losses

Evap

All NOx control measures

20300306

Internal Combustion Engines

Commercial/Institutional

Gasoline

Reciprocating:
Losses

Evap

All NOx control measures

20301006

Internal Combustion Engines

Commercial/Institutional

Liquified Petroleum Gas (LPG)

Reciprocating:
Losses

Evap

All NOx control measures

20200108

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Turbine: Evap Losses

All NOx control measures

20200109

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Turbine: Exhaust

NNSCRRBIC

20200208

Internal Combustion Engines

Industrial

Natural Gas

Turbine: Evap Losses

All NOx control measures

20200209

Internal Combustion Engines

Industrial

Natural Gas

Turbine: Exhaust

NNSCRRBIC2

20200908

Internal Combustion Engines

Industrial

Kerosene/Naphtha (Jet Fuel)

Turbine: Evap Losses

All NOx control measures

20200909

Internal Combustion Engines

Industrial

Kerosene/Naphtha (Jet Fuel)

Turbine: Exhaust

NNSCRRBGD

(continued)


-------
Table 5-9. Recommended Control Measure Deletions From SCCs in the CMDB (continued)

see

SCC Level 1

SCC Level 2

SCC Level 3

SCC Level 4

Control Measures
Recommended for Deletion

20201008

Internal Combustion Engines

Industrial

Liquified Petroleum Gas (LPG)

Turbine: Evap Losses

All NOx control measures

20201009

Internal Combustion Engines

Industrial

Liquified Petroleum Gas (LPG)

Turbine: Exhaust

NNSCRRBGD

20201011

Internal Combustion Engines

Industrial

Liquified Petroleum Gas (LPG)

Turbine

NNSCRRBGD

20201013

Internal Combustion Engines

Industrial

Liquified Petroleum Gas (LPG)

Turbine: Cogeneration

NNSCRRBGD

20300108

Internal Combustion Engines

Commercial/Institutional

Distillate Oil (Diesel)

Turbine: Evap Losses

All NOx control measures

20300109

Internal Combustion Engines

Commercial/Institutional

Distillate Oil (Diesel)

Turbine: Exhaust

NNSCRRBIC

20300208

Internal Combustion Engines

Commercial/Institutional

Natural Gas

Turbine: Evap Losses

All NOx control measures

20300209

Internal Combustion Engines

Commercial/Institutional

Natural Gas

Turbine: Exhaust

NNSCRRBIC2

20200105

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Reciprocating: Crankcase
Blowby

NNSCRWGTOL, NWTINGTOL

20200107

Internal Combustion Engines

Industrial

Distillate Oil (Diesel)

Reciprocating: Exhaust

NSCRWGTOL, NWTINGTOL

20200205

Internal Combustion Engines

Industrial

Natural Gas

Reciprocating: Crankcase
Blowby

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWTINGTNG

20200207

Internal Combustion Engines

Industrial

Natural Gas

Reciprocating: Exhaust

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWTINGTNG

20200252

Internal Combustion Engines

Industrial

Natural Gas

2-cycle Lean Burn

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWTINGTNG

20200253

Internal Combustion Engines

Industrial

Natural Gas

4-cycle Rich Burn

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWTINGTNG

20200254

Internal Combustion Engines

Industrial

Natural Gas

4-cycle Lean Burn

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWTINGTNG

20200255

Internal Combustion Engines

Industrial

Natural Gas

2-cycle Clean Burn

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWTINGTNG

(continued)


-------
Table 5-9. Recommended Control Measure Deletions From SCCs in the CMDB (continued)

see

SCC Level 1

SCC Level 2

SCC Level 3

SCC Level 4

Control Measures
Recommended for Deletion

20200256

Internal Combustion Engines

Industrial

Natural Gas

4-cycle Clean Burn

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWUNGTNG

20200905

Internal Combustion Engines

Industrial

Kerosene/Naphtha (Jet Fuel)

Reciprocating: Crankcase
Blowby

NSCRWGTJF, NWTINGTJF

20200907

Internal Combustion Engines

Industrial

Kerosene/Naphtha (Jet Fuel)

Reciprocating: Exhaust

NSCRWGTJF, NWTINGTJF

20300105

Internal Combustion Engines

Commercial/Institutional

Distillate Oil (Diesel)

Reciprocating: Crankcase
Blowby

NNSCRWGTOL, NWUNGTOL

20300107

Internal Combustion Engines

Commercial/Institutional

Distillate Oil (Diesel)

Reciprocating: Exhaust

NSCRWGTOL, NWUNGTOL

20300205

Internal Combustion Engines

Commercial/Institutional

Natural Gas

Reciprocating: Crankcase
Blowby

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWUNGTNG

20300207

Internal Combustion Engines

Commercial/Institutional

Natural Gas

Reciprocating: Exhaust

NLNBUGTNG, NSCRLGTNG,
NSCRSGTNG, NSCRWGTNG,
NSTINGTNG, NWUNGTNG


-------
Table 5-10. NOx Control Requirements for RICE in Pennsylvania

State

Source Category
Covered

NOx Control Level

Reference

Pennsylvania

153 tonNOx/season

>2400 hp: 3 g/bhp-hr (220 ppm)

IEPA 2007

Pennsylvania
(proposed
values)
[Assume
proposal was
201 lMar26]

General Permit—Natural
Gas Production and
Processing Facility, SI, ICE

Existing LB or RB, 100 to 1500 hp: 2 g/bhp-hr
New, Reconfigured, LB <100 hp: 2 g/bhp-hr
New, Reconfigured, LB 100 to 637 hp: 1 g/bhp-hr
New, Reconfigured, LB >637 hp: 0.5 g/bhp-hr

OTC 2012

Pennsylvania

(amended

2013Feb02)

Natural Gas Compression
and Processing, NG, SI,
ICE, includes facilities
with actual or potential
emissions <100 tpy NOx,
and <25 tpy NOx in 5
counties.

New, Reconfigured LB or RB, <100 hp: 2 g/bhp-hr
New, Reconfigured LB, 100 to 500 hp: 1 g/bhp-hr
New, Reconfigured LB, >500 hp: 0.5 g/bhp-hr
New, Reconfigured RB, 100 to 500 hp: 0.25 g/bhp-hr
New, Reconfigured RB, >500 hp: 0.2 g/bhp-hr

PA OKI'2013

Pennsylvania

Interstate Pollution
Transport Reduction,
Emission of NOx from
Stationary ICE

LB, >2400 hp: 3.0 g/bhp-hr
RICE, RB, >2400 hp: 1.5 g/bhp-hr

DE 2012

Table 5-11.

Characteristics of NOx

Kinissions and Controls for RICE



Engine Type and Uncontrolled
Size Emissions

Cost Effectiveness for
Emissions NOx Reductions

Reference

Lean burn







LB >500 hp

NA

SCR, stack test, 0.22 SCR: $71,000 to $60,000/ton
to 0.50 g/bhp-hr (for 500 to 4000 hp)

PADEP 2013,
p. 22

LB 100 to 500 hp 1 to 16.4 g/bhp-hr

NA SCR: >$42,000/ton

PADEP 2013,
p. 20

LB <100 hp

2 g/bhp-hr

2 g/bhp-hr SCR: >$48,000/ton

PADEP 2013,
p. 17

Rich burn







RB >500 hp

13 to 16 g/bhp-hr

NSCR: stack test, 0.02 NA
to 0.14 g/bhp-hr

PADEP 2013,
p. 28, 29

RB 100 to 500 hp 13 to 16.4 g/bhp-hr.

NA NSCR: $177/ton

PADEP 2013,
p. 25, 26

RB <100 hp

11.41 to 21.08 g/bhp-
hr

NSCR: <2 g/bhp-hr, at NSCR: <$650/ton for 100 hp
least 90% reduction NSCR. <$ j 200/ton for 50 hp

PADEP 2013,
p. 17

5-21


-------
APPENDIX A
AMMONIA REFORMERS

Copies of database tables showing all records for ammonia reformer controls,
highlighting revisions.

A-l


-------
Table A-l. CMDB Table 06 References (New)

Data Source

Description

AR-1

Indian Nations Council of Governments (INCOG), 2008: Indian Nations Council of Governments (INCOG), "Tulsa



Metropolitan Area 8-Hour Ozone Flex Plan: 2008 8-03 Flex Program," March 6. 2008. Downloaded from



htto://www.era.eov/ozoneadvance/i3dfs/Flex-Tulsa.i3df

t>

K>


-------
Table A-2. CMDB Table 01 Summary

cmname

Cm
Abbreviation

Pechan
Meas
Code

Major
Poll

Control
Technoloqv

Source Group

Sector

Class

Equip
Life

Nei Device
Code

Date
Reviewed

Data
Source

Months

Description

Low NOx
Burner;

Ammonia—NG-
Fired Reformers

NLNBUFRNG

N0561

NOx

Low NOx
Burner

Ammonia—NG-Fired
Reform ers

ptnonipm

Known

10

204|205

2013

AR-1 1186



Application: This control is the use of low NOx burner (LNB) technology to reduce NOx emissions. LNBs
reduce the amount of NOx created from reaction between fuel nitrogen and oxygen by lowering the
temperature of one combustion zone and reducing the amount of oxygen available in another.

This control is applicable to small (<1 ton NOx per OSD) ammonia production operations with natural gas-
fired reformers (SCC 30100306) and uncontrolled NOx emissions greater than 10 tons per year.

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-
rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply
excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air
LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air,
which acts as a heat sink to lower combustion temperatures (EPA, 2002).

Low NOx
Burner and Flue
Gas

Recirculation;
Ammonia—NG-
Fired Reformers

NLNBFFRNG

N0562

NOx

Low NOx
Burner and
Flue Gas
Recirculation

Ammonia—NG-Fired
Reform ers

ptnonipm

Known

10



2006

72|172|175|
179|186



Application: This control is the use of low NOx burner (LNB) technology and flue gas recirculation (FGR) to
reduce NOx emissions. LNBs reduce the amount of NOx created from reaction between fuel nitrogen and
oxygen by lowering the temperature of one combustion zone and reducing the amount of oxygen available
in another.

This control is applicable to small (<1 ton NOx per OSD) ammonia production operations with natural gas-
fired reformers (SCC 30100306) and uncontrolled NOx emissions greater than 10 tons per year.

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-
rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply
excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air
LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air,
which acts as a heat sink to lower combustion temperatures (EPA, 2002).

Low NOx
Burner and Flue
Gas

Recirculation;
Ammonia—Oil-
Fired Reformers

NLNBFFROL

N0572

NOx

Low NOx
Burner and
Flue Gas
Recirculation

Ammonia—Oil-Fired
Reform ers

ptnonipm

Known

10



2006





Application: This control is the use of low NOx burner (LNB) technology and flue gas recirculation (FGR) to
reduce NOx emissions. LNBs reduce the amount of NOx created from reaction between fuel nitrogen and
oxygen by lowering the temperature of one combustion zone and reducing the amount of oxygen available
in another.

This control is applicable to ammonia production operations with oil-fired reformers (SCC 30100307).

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-
rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply
excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air
LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air,
which acts as a heat sink to lower combustion temperatures (EPA, 2002).

Low NOx
Burner and Flue
Gas

Recirculation;
Ammonia Prod;
Feedstock
Desulfurization

NLNBFAPFD

N0622

NOx

Low NOx
Burner and
Flue Gas
Recirculation

Ammonia Prod;

Feedstock

Desulfurization

ptnonipm

Known

10



2006

72|172|175|
179| 185



Application: This control is the use of low NOx burner (LNB) technology and flue gas recirculation (FGR) to
reduce NOx emissions. LNBs reduce the amount of NOx created from reaction between fuel nitrogen and
oxygen by lowering the temperature of one combustion zone and reducing the amount of oxygen available
in another.

This control is applicable to small (<1 ton per OSD) feedstock desulfurization processes in ammonia
products operations (SCC 3Q1QQ3Q5) with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: It is assumed that the superheated steam needed to regenerate the activated carbon bed
used in the desulfurization process is the NOx source.

LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-rich
combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply excess
air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air LNBs
create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone. Staged-
fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air, which
acts as a heat sink to lower combustion temperatures (EPA, 2002).

Oxygen Trim
and Water
Injection;
Ammonia—NG-
Fired Reformers

NOTWIFRNG

N0563

NOx







Known

10



2006

72|172|175|
179| 184| 18
5



Application: This control is the use of OT + Wl to reduce NOx emissions

This control is applicable to small (<1 ton NOx per OSD) ammonia production operations
with natural gas-fired reformers (SCC 30100306) and uncontrolled NOx emissions greater
than 10 tons per year. This control is also applicable to miscellaneous combustion
emissions from ammonia production operations (SCC 30100399).

Discussion: Water is injected into the gas turbine, reducing the temperatures in the NOx-forming regions.
The water can be injected into the fuel, the combustion air or directly into the combustion chamber (ERG,
2000).

(continued)


-------
Table A-2. CMDB Table 01 Summary (continued)

cmname

Cm
Abbreviation

Pechan
Meas
Code

Major
Poll

Control
Technoloqv

Source Group

Sector

Class

Equip
Life

Nei Device
Code

Date
Reviewed

Data
Source

Months

Description

Selective
Catalytic
Reduction;
Ammonia—NG-
Fired Reformers

NSCRFRNG

N0564

NOx

Selective
Catalytic
Reduction

Ammonia—NG-Fired
Reform ers

ptnonipm

Known

20

139

2006

72|167|175|
179|224|22
5|226



Application: This control is the seieuive uaiaiyiic reduction of NOx through add-on controls. SCR controls
are post-combustion control technologies based on the chemical reduction of nitrogen oxides (NOx) into
molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal
efficiency, which allows the process to occur at lower temperatures.

Applies to natural-gas fired reformers involved in the production of ammonia (SCC 30100306) with
uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-
fired, combustion units for emission control since the early 1970s. SCR is typically implemented on units
requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA,

2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference
between SNCR and SCR is that SCR uses a metal-based catalyst to increase the rate of reaction (EPA,
2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The
reagent reacts selectively with the flue gas NOx within a specific temperature range and in the presence of
the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction
efficiency and the lower and broader temperature ranges. However, the decrease in reaction temperature
and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA,
2002). The cost increase is due to the large amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is
a gas at atmospheric pressure and normal temperatures. There are safety issues with the use of
anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous
ammonia is qenerallv transported and stored at a concentration of 29.4% ammonia in water.

Selective
Catalytic
Reduction;
Ammonia—Oil-
Fired Reformers

NSCRFROL

N0573

NOx

Selective
Catalytic
Reduction

Ammonia—Oil-Fired
Reform ers

ptnonipm

Known

20

139

2006





Application: This control is the selective catalytic reduction of NOx through add-on controls. SCR controls
are post-combustion control technologies based on the chemical reduction of nitrogen oxides (NOx) into
molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal
efficiency, which allows the process to occur at lower temperatures.

Applies to natural-gas fired reformers involved in the production of ammonia (SCC 30100306) with
uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-
fired, combustion units for emission control since the early 1970s. SCR is typically implemented on units
requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA,

2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference
between SNCR and SCR is that SCR uses a metal-based catalyst to increase the rate of reaction (EPA,
2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The
reagent reacts selectively with the flue gas NOx within a specific temperature range and in the presence of
the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction
efficiency and the lower and broader temperature ranges. However, the decrease in reaction temperature
and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA,
2002). The cost increase is due to the large amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is
a gas at atmospheric pressure and normal temperatures. There are safety issues with the use of
anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous
ammonia is generally transported and stored at a concentration of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support
structure. Most catalyst formulations contain additional compounds or sup-ports, providing thermal and
structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and
operational factors that affect the rate of reduction include: reaction temperature range; residence time
available in the optimum temperature range; degree of mixing between the injected reagent and the
combustion gases; uncontrolled NOx concentration level; molar ratio of injected reagent to uncontrolled
NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch;
catalyst deactivation; and catalyst management (EPA, 2001).

(continued)


-------
Table A-2. CMDB Table 01 Summary (continued)

cmname

Cm
Abbreviation

Pechan
Meas
Code

Major
Poll

Control
Technoloqv

Source Group

Sector

Class

Equip
Life

Nei Device
Code

Date
Reviewed

Data
Source

Months

Description

Selective Non-
Catalytic
Reduction—
Ammonia; NG-
Fired Reformers

NSNCRFRNG

N0565

NOx

Selective

Non-Catalytic

Reduction

Ammonia—NG-Fired
Reform ers

ptnonipm

Known

20

107

2006

72|172|175|
179| 185



Application: This control is the reduction of NOx emission through selective non-catalytic reduction add-on
controls. SNCR controls are post-combustion control technologies based on the chemical reduction of
nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20).

This control applies to small (<1 ton NOx per OSD) ammonia production natural gas fired reformers (SCC
30100306) with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: SNCR is the reduction of NOx in flue gas to N2 and water vapor. This reduction is done with a
nitrogen based reducing reagent, such as ammonia or urea. The reagent can react with a number of flue
gas components. However, the NOx reduction reaction is favored for a specific temperature range and in
the presence of oxygen (EPA, 2002).

Both ammonia and urea are used as reagents. The cost of the reagent represents a large part of the
annual costs of an SNCR system. Ammonia is generally less expensive than urea. However, the choice of
reagent is also based on physical properties and operational considerations (EPA, 2002).

Ammonia can be utilized in either aqueous or anhydrous form. Anhydrous ammonia is a gas at
atmospheric pressure and normal temperatures. There are safety issues with the use of anhydrous
ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is
generally transported and stored at a concentration of 29.4% ammonia in water.

Urea based systems have several advantages, including several safety aspects. Urea is a nontoxic, less
volatile liquid that can be stored and handled more safely than ammonia. Urea solution droplets can
penetrate farther into the flue gas when injected into the boiler, enhancing mixing (EPA, 2002). Because of
these advantages, urea is more commonly used than ammonia in large boiler applications.

Low NOx
Burner;

Ammonia—Oil-
Fired Reformers

NLNBUFROL

N0571

NOx

Low NOx
Burner

Ammonia—Oil-Fired
Reform ers

ptnonipm

Known

10

204|205

2006

72



Application: This control is the use of low NOx burner (LNB) technology to reduce NOx emissions. LNBs
reduce the amount of NOx created from reaction between fuel nitrogen and oxygen by lowering the
temperature of one combustion zone and reducing the amount of oxygen available in another.

This control is applicable to ammonia production operations with oil-fired reformers (SCC 30100307).

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-
rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply
excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air
LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air,
which acts as a heat sink to lower combustion temperatures (EPA, 2002).

Selective Non-
Catalytic
Reduction—
Ammonia; Oil-
Fired Reformers

NSNCRFROL

N0574

NOx

Selective

Non-Catalytic

Reduction

Ammonia—Oil-Fired
Reform ers

ptnonipm

Known

20

107

2006

72



Application: This control is the reduction of NOx emission through selective non-catalytic reduction add-on
controls. SNCR controls are post-combustion control technologies based on the chemical reduction of
nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20).

This control applies to ammonia production natural gas fired reformers (SCC 30100306) with uncontrolled
NOx emissions greater than 10 tons per year.

Discussion: SNCR is the reduction of NOx in flue gas to N2 and water vapor. This reduction is done with a
nitrogen based reducing reagent, such as ammonia or urea. The reagent can react with a number of flue
gas components. However, the NOx reduction reaction is favored for a specific temperature range and in
the presence of oxygen (EPA, 2002).

Both ammonia and urea are used as reagents. The cost of the reagent represents a large part of the
annual costs of an SNCR system. Ammonia is generally less expensive than urea. However, the choice of
reagent is also based on physical properties and operational considerations (EPA, 2002).

Ammonia can be utilized in either aqueous or anhydrous form. Anhydrous ammonia is a gas at
atmospheric pressure and normal temperatures. There are safety issues with the use of anhydrous
ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is
generally transported and stored at a concentration of 29.4% ammonia in water.

Urea based systems have several advantages, including several safety aspects. Urea is a nontoxic, less
volatile liquid that can be stored and handled more safely than ammonia. Urea solution droplets can
penetrate farther into the flue gas when injected into the boiler, enhancing mixing (EPA, 2002). Because of
these advantages, urea is more commonly used than ammonia in larqe boiler applications.

(continued)


-------
Table A-2. CMDB Table 01 Summary (continued)

cmname

Cm
Abbreviation

Pechan
Meas
Code

Major
Poll

Control
Technoloqv

Source Group

Sector

Class

Equip
Life

Nei Device
Code

Date
Reviewed

Data
Source

Months

Description

Low NOx

Burner;

Ammonia

Production;

Other Not

Classified

NLNBUAONC



NOx

Low NOx
Burner

Ammonia
Production—Other
Not Classified

ptnonipm

Known

10

204|205

2013

AR-11186



Application: This control is the use of low NOx burner (LNB) technology to reduce NOx emissions. LNBs
reduce the amount of NOx created from reaction between fuel nitrogen and oxygen by lowering the
temperature of one combustion zone and reducing the amount of oxygen available in another.

This control is applicable to miscellaneous combustion emissions from ammonia production operations
(SCC 30100399).

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-
rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply
excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air
LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air,
which acts as a heat sink to lower combustion temperatures (EPA, 2002)

Low NOx
Burner and Flue
Gas

Recirculation;
Ammonia
Production;
Other Not
Classified

NLNBFAONC



NOx

Low NOx
Burner and
Flue Gas
Recirculation

Ammonia
Production—Other
Not Classified

ptnonipm

Known

10



2013

72|172|175|
179|186



Application: This control is the use of low NOx burner (LNB) technology and flue gas
recirculation (FGR) to reduce NOx emissions. LNBs reduce the amount of NOx created
from reaction between fuel nitrogen and oxygen by lowering the temperature of one
combustion zone and reducing the amount of oxygen available in another.

This control is applicable to miscellaneous combustion emissions from ammonia production operations
(SCC 30100399).

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-
rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply
excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air
LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air,
which acts as a heat sink to lower combustion temperatures (EPA, 2002).

Selective Non-
Catalytic
Reduction—
Ammonia;
Ammonia
Production;
Other Not
Classified

NSNCRAONC



NOx

Selective

Non-Catalytic

Reduction

Ammonia
Production—Other
Not Classified

ptnonipm

Known

20

107

2013





Application: This control is the reduction of NOx emission through selective non-catalytic reduction add-on
controls. SNCR controls are post-combustion control technologies based on the chemical reduction of
nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20).

This control is applicable to miscellaneous combustion emissions from ammonia production operations
(SCC 30100399).

Discussion: SNCR is the reduction of NOx in flue gas to N2 and water vapor. This reduction is done with a
nitrogen based reducing reagent, such as ammonia or urea. The reagent can react with a number of flue
gas components. However, the NOx reduction reaction is favored for a specific temperature range and in
the presence of oxygen (EPA, 2002).

Both ammonia and urea are used as reagents. The cost of the reagent represents a large part of the
annual costs of an SNCR system. Ammonia is generally less expensive than urea. However, the choice of
reagent is also based on physical properties and operational considerations (EPA, 2002).

Ammonia can be utilized in either aqueous or anhydrous form. Anhydrous ammonia is a gas at
atmospheric pressure and normal temperatures. There are safety issues with the use of anhydrous
ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is
generally transported and stored at a concentration of 29.4% ammonia in water.

Urea based systems have several advantages, including several safety aspects. Urea is a nontoxic, less
volatile liquid that can be stored and handled more safely than ammonia. Urea solution droplets can
penetrate farther into the flue gas when injected into the boiler, enhancing mixing (EPA, 2002). Because of
these advantages, urea is more commonly used than ammonia in large boiler applications.

Oxygen Trim

and Water

Injection;

Ammonia

Production;

Other Not

Classified

NOTWIAONC



NOx

Oxygen Trim
and Water
Injection

Ammonia
Production—Other

Nnt ClaccifipiH

ptnonipm

Known

10



2013

72|172|175|
179| 184| 18
5



Application: This control is the use of OT + Wl to reduce NOx emissions

This control is applicable to miscellaneous combustion emissions from ammonia
production operations (SCC 30100399).

Discussion: Water is injected into the gas turbine, reducing the temperatures in the NOx-forming regions.
The water can be injected into the fuel, the combustion air or directly into the combustion chamber (ERG,
2000).

(continued)


-------
Table A-2. CMDB Table 01 Summary (continued)

cmname

Cm
Abbreviation

Pechan
Meas
Code

Major
Poll

Control
Technoloqv

Source Group

Sector

Class

Equip
Life

Nei Device
Code

Date
Reviewed

Data
Source

Months

Description

Selective

Catalytic

Reduction;

Ammonia

Production;

Other Not

Classified

NSCRAONC



NOx

Selective
Catalytic
Reduction

Ammonia
Production—Other
Not Classified

ptnonipm

Known

20

139

2013

72|167|175|
179|224|22
5|226



Application: This control is the seieuuve uaiaiyiic reduction of NOx through add-on controls. SCR controls
are post-combustion control technologies based on the chemical reduction of nitrogen oxides (NOx) into
molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal
efficiency, which allows the process to occur at lower temperatures.

This control is applicable to miscellaneous combustion emissions from ammonia production operations
(SCC 30100399).

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-
fired, combustion units for emission control since the early 1970s. SCR is typically implemented on units
requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA,

2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference
between SNCR and SCR is that SCR uses a metal-based catalyst to increase the rate of reaction (EPA,
2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The
reagent reacts selectively with the flue gas NOx within a specific temperature range and in the presence of
the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction
efficiency and the lower and broader temperature ranges. However, the decrease in reaction temperature
and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA,
2002). The cost increase is due to the large amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent.

Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There
are safety issues with the use of anhydrous ammonia, as it must be transported and
stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and
stored at a concentration of 29.4% ammonia in water.

t>


-------
Table A-3. CMDB Table 02 Efficiencies

cmabbreviation

Pollutant

Locale

Effective
Date

existingmeasureabbr

neiexistingdevcode

minemissions

maxemissions

controlefficiency

costyear

costperton

ruleeff

rulepen

equationtype

caprecfactor

discountrate

capannratio

incrementalcpt

Details

NLNBFAPFD

NOx







0

0

365

60

1990

2560

100

100

cpton

0.1424



5.9

2470

Applied to small source types

NLNBFAPFD

NOx







0

365



60

1990

590

100

100

cpton

0.1424



7.5

280

Applied to large source types

NLNBFFRNG

NOx







0

0

365

60

1990

2560

100

100

cpton

0.1424



5.9

2470

Applied to small source types

NLNBFFRNG

NOx







0

365



60

1990

590

100

100

cpton

0.1424



7.5

280

Applied to large source types

NLNBFFROL

NOx







0

0

365

60

1990

1120

100

100

cpton

0.1424



5.9

1080

Applied to small source types

NLNBFFROL

NOx







0

365



60

1990

390

100

100

cpton

0.1424



7.5

190

Applied to large source types

NLNBUFROL

NOx







0

0

365

50

1990

400

100

100

cpton

0.1424



5.5



Applied to small source types

NLNBUFROL

NOx







0

365



50

1990

430

100

100

cpton

0.1424



5.5



Applied to large source types

NOTWIFRNG

NOx







0

0

365

65

1990

680

100

100

cpton

0.1424



2.9



Applied to small source types

NOTWIFRNG

NOx







0

365



65

1990

320

100

100

cpton

0.1424



2.9



Applied to large source types

NSCRFRNG

NOx







0

0

365

90

1999

2366

100

100

cpton

0.0944



10



Applied to small source types

NSCRFRNG

NOx







0

365



90

1999

2366

100

100

cpton

0.0944



9.6



Applied to large source types

NSCRFROL

NOx







0

0

365

80

1990

1480

100

100

cpton

0.0944



10

1910

Applied to small source types

NSCRFROL

NOx







0

365



80

1990

810

100

100

cpton

0.0944



9.6

940

Applied to large source types

NSNCRFRNG

NOx







0

0

365

50

1990

3870

100

100

cpton

0.0944



9.4

2900

Applied to small source types

NSNCRFRNG

NOx







0

365



50

1990

1570

100

100

cpton

0.0944



8.2

840

Applied to large source types

NSNCRFROL

NOx







0

0

365

50

1990

2580

100

100

cpton

0.0944



9.4

1940

Applied to small source types

NSNCRFROL

NOx







0

365



50

1990

1050

100

100

cpton

0.0944



8.2

560

Applied to large source types

NLNBUFRNG

NOx







0

0

365

50

1990

820

100

100

cpton

0.1424



5.5



Applied to small source types; no new
information was available for small sources
during 2013 update

NLNBUFRNG

NOx







0

365



50

2008

800

100

100

cpton

0.1424



5.9



Applied to large source types; equipment
life of 10 years and 7% interest


-------
APPENDIX B
COMBUSTION TURBINES

Copies of the database tables for showing all records for Combustion Turbines NOx
controls are provided. Changes are highlighted in red font.

-	Table B-l. CMDB Table OlSummary

-	Table B-2. CMDB Table 02_Efficiencies

-	Table B-3. CMDB Table 04_Equations

-	Table B-4. Additional CMDB Table 06 References

B-l


-------
Table B-l. CMDB Table OlSummary

cmname

cmabbreviation

pechanmea
scode

majorpoll

controltechnologv

sourcegroup

sector

class

equiplife

ll'iTHlli'

dntcrcviewed

datasource

months

Dry Low NOx
Combustion; Gas
Turbines—Natural Gas

NDLNCGTNG

N0243

NOx

Dry Low NOx
Combustion

Gas Turbines—
Natural Gas

ptnonipm

Known

15





72 172 175 179 22
3 CT-2 CT-6



SCR + Dry Low NOx
Combustion; Gas
Turbines—Natural Gas

NSCRDGTNG

N0244

NOx

SCR + DLN
Combustion

Gas Turbines—
Natural Gas

ptnonipm

Known

15



2013

72 172 175 179 22
3 224 CT-2 CT-
3 CT-4 CT-6 CT-8



Selective Catalytic
Reduction and Steam
Injecti; Gas Turbines—
Natural Gas

NSCRSGTNG

N0245

NOx

Selective Catalytic
Reduction and
Steam Injection

Gas Turbines—
Natural Gas

ptnonipm

Known

15



2013

72 172 175 179 22
3 224 CT-2 CT-3



Selective Catalytic
Reduction and Water
Injecti; Gas Turbines—
Jet Fuel

NSCRWGTJF

N0502

NOx

Selective Catalytic
Reduction and
Water Injection

Gas Turbines—
Jet Fuel

ptnonipm

Known





2013

72 172 175 179 22
3 CT-2 CT-7



Selective Catalytic
Reduction and Water
Injecti; Gas Turbines—
Natural Gas

NSCRWGTNG

N0246

NOx

Selective Catalytic
Reduction and
Water Injection

Gas Turbines—
Natural Gas

ptnonipm

Known

15



2013

72 172 175 179 22
3 224 CT-2 CT-
3 CT-8



Selective Catalytic
Reduction and Water
Injecti; Gas Turbines—
Oil

NSCRWGTOL

N0232

NOx

Selective Catalytic
Reduction and
Water Injection

Gas Turbines—
Oil

ptnc nil 'iii

TInown

15



2013

72 172 175 179 22
3 224 CT-2 CT-7



Steam Injection; Gas
Turbines—Natural Gas

NSTINGTNG

N0242

NOx

Steam Injection

Gas Turbines—
Natural Gas

ptnonipm

Known

15



2013

72 172 175 184 22
3 CT-2



Water Injection; Gas
Turbines—Jet Fuel

NWTINGTJF

N0501

NOx

Water Injection

Gas Turbines—
Jet Fuel

ptnonipm

Known

15



2013

72 172 175 184 22
3 CT-2



Water Injection; Gas
Turbines—Natural Gas

NWTINGTNG

N0241

NOx

Water Injection

Gas Turbines—
Natural Gas

ptnonipm

Known

15



2013

72 172 175 184 22
3 CT-2



Water Injection; Gas
Turbines—Oil

NWTINGTOL

N0231

NOx

Water Injection

Gas Turbines—

Oil

ptnonipm

Known

15



2013

72 172 175 184 22
3 CT-2



Catalytic Combustion;
Gas Turbine—Natural

Gas

NCATCGTNG

N/A

NOx

Catalytic
Combustion

Gas Turbines—
Natural Gas

ptnonipm

Emerging

15



2013

CT-1 CT-2



EMx and Dry Low NOx
Combustion; Gas
Turbines—Natural Gas

NEMXDGTNG

N/A

NOx

EMx and Dry Low
NOx Combustion

Gas Turbines—
Natural Gas

ptnonipm

Emerging

15



2013

CT-1 CT-2 CT-
3 CT-4 CT-5



EMx and Water
Injection; Gas
Turbines—Natural Gas

NEMXWGTNG

N/A

N( )x

EMx and Water
ink1' tion

Gas Turbines—
Natural Gas

ptnonipm

Emerging

15



2013

CT-1 CT-3



*For ease in reading this table, 1 lie Descnpiimi field is included on separate pages.


-------
Table B-l. CMDB Table OlSummary (continued)

cmabbreviation

Description

NDLNCGTNG

Application: This control is the use of dry low NOx combustion (DLN) technology to reduce NOx emissions. DLN combustion reduces the amount of NOx created from reaction
between fuel nitrogen and oxygen by lowering the temperature of one combustion zone and reducing the amount of oxygen available in another.

This control applies to large (83.3 MW to 161 MW) natural gas fired turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-rich combustion and one at a lower temperature. Staging techniques are
usually used by LNB to supply excess air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air LNBs create a fuel-rich reducing primary
combustion zone and a fuel-lean secondary combustion zone. Staged-fiiel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air, which acts as
a heat sink to lower combustion temperatures (EPA, 2002).

NSCRDGTNG

Application: This control is the selective catalytic reduction of NOx through add-on controls. SCR controls are post-combustion control technologies based on the chemical
reduction of nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal efficiency, which allows the
process to occur at lower temperatures.

This control applies to natural gas fired turbines with NOx emissions greater than 10 Ions per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil foel-fired, combustion units for emission control since the early 1970s. SCR is
typically implemented on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the
rate of reaction (EPA, 2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a
specific temperature range and in the presence of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the
decrease in reaction temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large
amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety
issues with the use of anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration
of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support structure. Most catalyst formulations contain additional compounds or sup-
ports, providing thermal and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction
temperature range; residence time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx
concentration level; molar ratio of injected reagent to uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst
deactivation; and catalyst management (EPA, 2001).

NSCRSGTNG

Application: This control is the selective catalytic reduction of NOx through add-on controls. SCR controls are post-combustion control technologies based on the chemical
reduction of nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal efficiency, which allows the
process to occur at lower temperatures.

This control applies to natural gas fired turbines with NOx emissions greater than 10 tons per year.

(continued)


-------
Table B-l. CMDB Table OlSummary (continued)

cmabbreviation

Description

NSCRSGTNG
(cont.)

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is
typically implemented on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002),

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the
rate of reaction (EPA, 2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a
specific temperature range and in the presence of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the
decrease in reaction temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large
amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety
issues with the use of anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration
of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support structure. Most catalyst formulations contain additional compounds or sup-
ports, providing thermal and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction
temperature range; residence time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx
concentration level; molar ratio of injected reagent to uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst
deactivation; and catalyst management (EPA, 2001).

NSCRWGTJF

Application: This control is the selective catalytic reduction of NOx through add-on controls in combination with water injection. SCR controls are post-combustion control
technologies based on the chemical reduction of nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx
removal efficiency, which allows the process to occur at lower temperatures.

This control applies to jet fuel-fired turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is
typically implemented on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the
rate of reaction (EPA, 2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a
specific temperature range and in the presence of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in luo advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the
decrease in reaction temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large
amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety
issues with the use of anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration
of 29.4% ammonia in water.

(continued)


-------
Table B-l. CMDB Table OlSummary (continued)

cmabbreviation

Description

NSCRWGTJF
(cont.)

Today, catalyst formulations include single component, multi-component, or active phase with a support structure. Most calalysl formulations contain additional compounds or sup-
ports, providing thermal and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction
temperature range; residence time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx
concentration level; molar ratio of injected reagent to uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst
deactivation; and catalyst management (EPA, 2001).

NSCRWGTNG

Application: This control is the selective catalytic reduction of NOx through add-on controls in combination with water injection. SCR controls are post-combustion control
technologies based on the chemical reduction of nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx
removal efficiency, which allows the process to occur at lower temperatures.

This control applies to natural gas-fired gas turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is
typically implemented on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the
rate of reaction (EPA, 2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a
specific temperature range and in the presence of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the
decrease in reaction temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large
amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety
issues with the use of anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration
of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support structure. Most catalyst formulations contain additional compounds or sup-
ports, providing thermal and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction
temperature range; residence time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx
concentration level; molar ratio of injected reagent to uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst
deactivation; and catalyst management (EPA, 2001).

NSCRWGTOL

Application: This control is the selective catalytic reduction of NOx through add-on controls in combination with water injection. SCR controls are post-combustion control
technologies based on the chemical reduction of nitrogen oxides (NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx
removal efficiency, which allows the process to occur at lower temperatures.

This control applies to oil-fired turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is
typically implemented on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

(continued)


-------
Table B-l. CMDB Table OlSummary (continued)

cmabbreviation

Description

NSCRWGTOL
(cont.)

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is lhal SCR uses a metal-based catalyst to increase the
rate of reaction (EPA, 2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a
specific temperature range and in the presence of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the
decrease in reaction temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large
amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety
issues with the use of anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration
of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support structure. Most catalyst formulations contain additional compounds or sup-
ports, providing thermal and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction
temperature range; residence time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx
concentration level; molar ratio of injected reagent to uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst
deactivation; and catalyst management (EPA, 2001).

NSTINGTNG

Application: This control is the use of steam injection to reduce NOx emissions.

This control applies to small (3.3 MW to 34.4MW) natural gas-fired gas turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Steam is injected into the gas turbine, reducing the temperatures in the NOx-forming regions. The steam can be injected into the fuel, the combustion air or directly into
the combustion chamber (ERG, 2000).

NWTINGTJF

Application: This control is the use of water injection to reduce NOx emissions.

This control applies to small (3.3 MW to 34.4MW) jet fuel-fired turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Water is injected into the gas turbine, reducing the temperatures in the NOx-forming regions. The water can be injected into the fuel, the combustion air or directly into
the combustion chamber (ERG, 2000).

NWTINGTNG

Application: This control is the use of water injection to reduce NOx emissions.

This control applies to small (3.3 MW to 34.4MW) natural gas-fired gas turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Water is injected into the gas turbine, reducing the temperatures in the NOx-forming regions. The water can be injected into the fuel, the combustion air or directly into
the combustion chamber (ERG, 2000).

NWTINGTOL

Application: This control is the use of water injection to reduce NOx emissions.

This control applies to small (3.3 MW to 34.4MW) oil-fired turbines with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Water is injected into the gas turbine, reducing the temperatures in the NOx-forming regions. The water can be injected into the fuel, the combustion air or directly into
the combustion chamber (ERG, 2000).

(continued)


-------
Table B-l. CMDB Table OlSummary (continued)

cmabbrevlation

Description

NCATCGTNG

Application: This control is the use of catalytic combustion to reduce NOx emissions. Catalytic combustors reduce the amount of NOx created by oxidizing fuel at lower
temperatures (and without a flame) than in conventional combustors. Catalytic combustion uses a catalytic bed to oxidize a lean air fuel mixture within a combustor instead of
burning with a flame. The fuel and air mixture oxidizes at lower temperatures than in a conventional combustor, producing less NOx.

Currently installed only on a few 1.4 MW combustion turbines, and commercially available for turbines rated up to 10 MW (CT-1).

NEMXDGTNG

Application: This control is the use of EMx in combination with dry low NOx combustion. EMx is a post-combustion catalytic oxidation and absorption technology that uses a two-
stage catalyst/absorber system for the control of NOx as well as CO, VOC, and optionally SOx. A coated catalyst oxidizes NO to N02, CO to C02, and VOC to C02 and water.
The N02 is then absorbed onto the catalyst surface where it is chemically converted to and stored as potassium nitrates and nitrites. A proprietary regeneration gas is periodically
passed through the catalyst to desorb the N02 from the catalyst and reduce it to elemental nitrogen (N2). EMx has been successfully demonstrated on several small combustion
turbine projects up to 45 MW. The manufacturer has claimed that EMx can be effectively scaled up to larger turbines (CT-1).

Cost estimates for DLN combustion in 2008 dollars are not available. Thus, the total system cost in this analysis in 2008 dollars was developed from 1999 cost estimates for DLN
combustion that were escalated to 2008 dollars and added to the available 2008 estimate for the EMx system.

NEMXWGTNG

Application: This control is the use of EMx in combination with water injection.

Cost estimates for water injection in 2008 dollars are not available. Thus, the total system cos1 in this analysis in 2008 dollars was developed from 1999 cost estimates for water
injection that were escalated to 2008 dollars and added to the available 2008 estimate for Ihe I'.M\ system.

td


-------
Table B-2. CMDB Table 02 Efficiencies

cmabbreviation

pollutant

tj
©

Effective Date ||

existingmeasureabbr ||

neiexistingdevcode ||

minemissions

maxemissions

controlefficiency

costyear

costperton

ruleeff

rulepen

equationtype

caprecfactor

iliscountrate ||

capannratio

£
E

(J

e

details

NWTINGTNG

NOx







0

0

365

72

1999

1790

100

100

cpton

0.109::



3.1



Applied to small source types (<34.4 MW),
uncontrolled emissions <365 tpy

NWTINGTNG

NOx







0

365



72

1999

1000

100

100

cpton

0.109b



2.4



Applied to small source types (<34.4 MW),
uncontrolled emissions >365 tpy

NWTINGTNG

NOx







0

365



72

1999

730

100

100

cpton

0.1098



1.6



Applied to large source types

NSCRWGTNG

NOx







0

0

365

94

1999

2790

100

100

cpton

0.1098



3

5840

Applied to small source types (3 to 26 MW),
uncontrolled emissions <365 tpy.

NSCRWGTNG

NOx







0

365



94

1999

1370

100

100

cpton

0.1098



2.9

3130

Applied to small source types (3 to 26 MW),
uncontrolled emissions >365 tpy.

NSCRWGTNG

NOx







0

365



94

1999

1070

100

100

cpton

0.1098



1.5

1690

Applied to large source types (~80 to 160 MW)

NSCRWGTNG

NOx







0

365



98

2008

1960

100

100

cpton

0.1098



2.5

3170

Applied to large source types (~50 to 180 MW), 1999
costs for WI assumed to be the same as 1990 costs in
the 1993 ACT based on data in ref CT-2 that showed
the costs were essentially the same for NG-fired units.
1999 WI capital and indirect annual costs were
escalated to 2008 dollars using ratio of 2008 to 1999
CEP cost indexes, direct annual costs for WI were
assumed to be the same in 2008 as in 1999, and
resulting 2008 costs were added to the 2008 SCR costs
from ref CT-3.

NEMXWGTNG

NOx







0

365



99



2960

100

100

cpton

0.1098



2.9

7120

Applied to large source types (50 to 180 MW); WI costs
estimated using the same procedure as for
NSCRWGTNG applied to large sources.

NSTINGTNG

NOx







0

0



80



1690

100

100

cpton

0.1098



3.5



Applied to small source types (3 to 26.3 MW),
uncontrolled emissions <365 tpy, 1999 costs for SI
assumed to be the same as 1990 costs in the 1993 ACT
based on data in ref CT-2 that showed WI costs were
essentially the same for NG-fired units (assumed same
pattern holds for steam injection).

NSTINGTNG

NOx













80



820

100

100

cpton

0.1098



3.5



Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy, 1999 costs for SI
assumed to be the same as 1990 costs in the 1993 ACT
based on data in ref CT-2 that showed WI costs were
essentially the same for NG-fired units (assumed same
pattern holds for steam injection).

NSTINGTNG

NOx







o





80

1999

500

100

100

cpton

0.1098



3.0



Applied to large source types (~80 to 160 MW), 1999
costs for SI assumed to be the same as 1990 costs in the
1993 ACT based on data in ref CT-2 that showed WI
costs were essentially the same for NG-fired units
(assumed same pattern holds for steam injection).

NSCRSGTNG

NOx







0

0

365

95

1999

2570

100

100

cpton

0.1098



3.3

5550

Applied to small source types (3 to 26 MW),
uncontrolled emissions <365 tpy.


-------
(continued)

Table B-2. CMDB Table 02_Efficiencies (continued)

cmabbreviation

pollutant

tj
©

Effective Date ||

existingmeasureabbr ||

neiexistingdevcode ||

minemissions

maxemissions

>¦.
tj

c

*5

E

"©
•—

c

©
u

costyear

costperton

ruleeff

rulepen

equationtype

caprecfactor

«
•—

c

3

capannratio

<3

c
E

u

e

details

NSCRSGTNG

NOx







0

365



95

1999

1380

100

100

cpton

0.109::



3.1

2870

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy.

NSCRSGTNG

NOx







0

365



95

1999

570

100

100

cpton

0.1098



2.7

1810

Applied to large source types (~80 to 160 MW)

NSCRGYNG

NOx







0

365



95

2008

1420

100

100

cpton

0.1098



3.9

3170

Applied to large source types (50 to 180 MW)

NDLNCGTNG

NOx







0

0

365

84

1999

300

100

100

rrt^n

0.1098



5

S4&

Applied to small source types

NDLNCGTNG

NOx







0

365



84

1999

130

100

100



0.1098



7.4



Applied to large source types

NSCRDGTNG

NOx







0

0

365

94

1999

1800

100

100



0.1098



2.9

11900

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions <365 tpy.

NSCRDGTNG

NOx







0

365



94

1999

990

100

100

cpton

0.1098



3.6

6320

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy.

NSCRDGTNG

NOx







0

365



94

1999

390

100

100

cpton

0.1098



4.2

3340

Applied to large source types (~160 MW)

NSCRDGTNG

NOx







0

365





2007















18900

Applied to small source types (up to 40 MW,
uncontrolled emissions <365 tpy)

NSCRDGTNG

NOx







0



365



2007















7510

Applied to small source types (up to 40 MW,
uncontrolled emissions >365 tpy)

NSCRDGTNG

NOx







0

365



94

2008

1040

100

100

cpton

0.1098



4.6

5560

Applied to large source types (~50 to 180 MW), 1999
costs for DLN were estimated based on data in ref CT-
2. Escalated these costs to 2008 dollars using ratio of

2008 to 1999 CEP cost indexes and added to the 2008
SCR costs from ref CT-3.

NEMXDGTNG

NOx











365



1999

2860













14940

Applied to small source types (<26 MW), uncontrolled
emissions <365 tpy

NEMXDGTNG

NOx









365





1999

1720













10270

Applied to small source types (<26 MW), uncontrolled
emissions >365 tpy

NEMXDGTNG

NOx









365





1999

840













6600

Applied to large source types (170 MW), uncontrolled
emissions >365 tpy

NEMXDGTNG

NOx







0



365





















Applied to small source types

NEMXDGTNG

NOx















2008

2040

100

100

cpton

0.1098



4.1

12370

Applied to large source types (50 to 180 MW); DLN
costs estimated in 1999 dollars were escalated to 2008
dollars using the CEPCI, except parts and repair costs
were assumed to be the same in 2008 as in 1999.

NCATCGTNG

NOx















1999

920

100

100

cpton

0.1098



1.7

4760

Applied to small source types (3 to 26 MW),
uncontrolled emissions <365 tpy.

NCATCGTNG

NOx







o



365

98

1999

670

100

100

cpton

0.1098



1.2

2580

Applied to small source types (3 to 26 MW),
uncontrolled emissions >365 tpy.

NCATCGTNG

NOx







o

365



98

1999

370

100

100

cpton

0.1098



0.7

2200

Applied to large source types (~170 MW)

NWTINGTOL

NOx







o

0

365

68

1999

1630

100

100

cpton

0.1098



3.0



Applied to small source types (3 to 26.3 MW),
uncontrolled emissions <365 tpy, 1999 costs assumed to
be the same as 1990 costs in the 1993 ACT based on
data in ref CT-2 that showed the costs were essentially
the same for NG-fired units.


-------
(continued)

Table B-2. CMDB Table 02_Efficiencies (continued)

cmabbreviation

pollutant

tj
©

Effective Date ||

existingmeasureabbr ||

neiexistingdevcode ||

minemissions

maxemissions

>¦.
tj

c

*5

E

"©
•—

c

©
u

costyear

costperton

ruleeff

rulepen

equationtype

caprecfactor

«
•—

c

3

capannratio

<3

c
E

u

e

details

NWTINGTOL

NOx







0

365



68

1999

960

100

100

cpton

0.109-



1.8



Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy, 1999 costs assumed to
be the same as 1990 costs in the 1993 ACT based on
data in ref CT-2 that showed the costs were essentially
the same for NG-fired units.

NWTINGTOL

NOx







0

365



68

1999

650

100

100

cpton

0.1098



1.6



Applied to large source types (—83 MW), uncontrolled
emissions >365 tpy, 1999 costs assumed to be the same
as 1990 costs in the 1993 ACT based on data in ref CT-
2 that showed the costs were essentially the same for
NG-fired units.

NSCRWGTOL

NOx







0

0

365

90

1990

3190

100

100

cpton

0.1098



2.9

7620

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions <365 tpy.

NSCRWGTOL

NOx







0

365



90

1990

1320

100

100

cpton

0.1098



2.3

2450

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy.

NSCRWGTOL

NOx







0

365



97

2004

1560

100

100

cpton

0.1098



2.3

4790

Applied to large source types (—83 MW), uncontrolled
emissions >365 tpy, 1999 costs for WI assumed to be
the same as 1990 costs in the 1993 ACT based on data
in ref CT-2 that showed the costs were essentially the
same for NG-fired units. Escalated these costs to 2004
dollars using ratio of 2004 to 1999 CEP cost indexes
and added to the 2004 SCR costs from ref CT-7.
Control efficiency based on data from analysis for one
unit (ref CT-7).

NWTINGTJF

NOx







0

0

365

68

1999

1630

100

100

cpton

0.1098



3.0



Applied to small source types (3 to 26.3 MW),
uncontrolled emissions <365 tpy, costs assumed to be
the same as for oil-fired turbines.

NWTINGTJF

NOx







0

365



68

1999

960

100

100

cpton

0.1098



1.8



Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy, costs assumed to be
the same as for oil-fired turbines.

NWTINGTJF

NOx







o





68

1999

650

100

100

cpton

0.1098



1.6



Applied to large source types (—83 MW), uncontrolled
emissions >365 tpy, costs and control efficiency
assumed to be the same as for oil-fired turbines.

NSCRWGTJF

NOx







o

o



90

1990

3190

100

100

cpton

0.1098



2.9

7620

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions <365 tpy, costs assumed to be
same as for oil-fired turbines.

NSCRWGTJF

NOx







o





90

1990

1320

100

100

cpton

0.1098



2.3

2450

Applied to small source types (3 to 26.3 MW),
uncontrolled emissions >365 tpy, costs assumed to be
same as for oil-fired turbines.

NSCRWGTJF

NOx







0

365



97

2004

1560

100

100

cpton

0.1098



2.3

4790

Applied to large source types (—83 MW), uncontrolled
emissions >365 tpy, costs and control efficiency
assumed to be same as for oil-fired turbines).


-------
td


-------
Table B-3. CMDB Table 04_Equationsa

cmabbreviation

cmeqntype

pollutant

costyear

varl

var2

var3

var4

var5

\arfi

var7

var8

var9

varlO

NWTINGTNG

Type 2

NOx

1999

27665

0.69

3700.2

0.95

27665

u.oy

3700.2

0.95





NSCRWGTNG

Type 2

NOx

1999

62962

0.66

8590

0.87

37193

0.63

12065

0.64





NSCRWGTNG

Type 2

NOx

2007









210883

0.46









NSCRWGTNG

Type"L"

NOx

2007













1893.8

185570





NSCRWGTNG

Type 2

NOx

2008

34533

0.85

7236

0.94

10323

0.96

3106

0.94





NEMXWGTNG

Type 2

NOx

2008

200894

0.68

19215

0.86

160409

0.67

20174

0.78





NSHNGTNG

Type 2

NOx

1999

43092

0.66

7282.3

0.76

43092

0.66

7282.3

0.76





NSCRSGTNG

Type 2

NOx

1999

72169

0.66

17551

0.72

37193

0.63

12065

0.64





NSCRSGTNG

Type 2

NOx

2008

46492

0.82

9434.1

0.86

10323

0.96

3106

0.94





NDLNCGTNG

Type 2

NOx

1999





676.37

0.96





676.37

0.96





NDLNCGTNG

Type"L"

NOx

1999

2860.6

25427





2860.6

25427









NSCRDGTNG

Type 2

NOx

1999

24854

0.79

12725

0.69

37193

0.63

12065

0.64





NSCRDGTNG

Type 2

NOx

2007

187647

0.54





210883

0.46









NSCRDGTNG

Type"L"

NOx

2007





2782

167494





1893.8

185570





NSCRDGTNG

Type 2

NOx

2008

14785

0.97

5250.8

0.9

10323

0.96

3106.1

0.94





NEMXDGTNG

Type 2

NOx

1999

58237

0.78

15004

0.78

65163

0.72

13702

0.76





NEMXDGTNG

Type 2

NOx

2008

129611

0.74

23051

0.78

160409

0.67

20174

0.78





NCATCGTNG

Type 2

NOx

1999

20668

0.57

4254.2

0.82













NCATCGTNG

Type"L"

NOx

1999









N/A

N/A

743.2

54105





NWTINGTOL

Type 2

NOx

1999

42533

0.6

6776.7

0.8

42533

0.6

6776.7

0.8





NSCRWGTOL

Type 2

NOx

1990

94337

0.63

25914

0.7













NSCRWGTOL

Type"L"

NOx

1999









4868.5

349694

1546.1

139203





aType "L" is a linear equation; variables arc the slope and intercept. No incremental TCI for NCATCGTNG relative to DLN because the capital costs for
catalytic combustion are lower than the capital costs for DLN for all but the smallest turbines. The underlying data for 2008 costs for SCR and EMx are for
large turbines (50 MW to 180 MW). The underlying data for 2007 costs are for 1 MW to 40 MW turbines.


-------
Table B-4. Additional CMDB Table 06 References

Data Source

Description

CT-1

Bay Area Air Quality Management District, 2010. Preliminary Determination of Compliance. Marsh Landing Generating Station. March 2010.
Available at: http://www.energy.ca.gov/sitingcases/marshlanding/documents/other/2010-03-24_Bay_Arca_AQMD_PDOC.pdf

CT-2

Onsite Sycom Energy Corporation, 1999. "Cost Analysis of NOx Control Alternatives for Stationary Gas Turbines." Prepared for U.S.
Department of Energy. Environmental Programs Chicago Operations Office. November 5, 1999. Available at:
https://wwwl.eere.energy.gov/manufacturing/distributedenergy/pdfs/gas_turbines_nox_cost_analysis.pdf

CT-3

EmeraChem Power, 2008. Attachment in email from Jeff Valmus. EmcraChcm Power, to WcymanLee, BAAQMD. Request for EMx Cost
Information. September 8, 2008. Available at:

http://www.baaqmd.gOv/~/media/Files/Engineering/Public%20Noticcs/2010/18404/Foolnolcs/EMx%20BACT%20economic%20analysis%20f
inal09072008.ashx

CT-4

CH2MHill, 2002. Walnut Energy Center Application for Certification." Prepared for California Energy Commission. November 2002.
Available at: www.energy.ca.gov/sitingcases/turlock/documcnls/applicanl_filcs/volumc_2/App_08.01E_Eval_Control.pdf.

CT-5

CARB, 2004. California Environmental Protection Agency. Air Resources Board. Report to the Legislature. Gas-Fired Power Plant NOx
Emission Controls and Related Environmental Impacts. Stationary Source Division. May 2004. Available at:
http://www.arb.ca.gov/research/apr/reports/12069.pdf

CT-6

Resource Dynamics Corporation, 2001. "Assessment of Distributed Generation Technology Applications." Prepared for Maine Public Utilities
Commission. February 2001. Available at: http://www.dislribulcd-gcncralion.com/Library/Maine.pdf

CT-7

Florida Municipal Power Agency. 2004. Chapters 3 and 4 of PSD BACT analysis for Stock Island facility in Key West, Florida. Available at
http://www.dep.state.fl.us/air/cmission/conslruclion/slockisland/BasisofBACT.pdf and
http://www.dep.state.fl.us/air/cmission/construclion/slockisland/NOxBACT.pdf

CT-8

Energy and Environmental Analysis ( An ICF International Company), 2008. Technology Characterization: Gas Turbines. Prepared for
Environmental Protection Agency Climate Protection Partnership Division. December 2008. Available at:
http://www.epa.gov/chp/documcnls/calalog_chplcch_gas_lurbines.pdf


-------
APPENDIX C
GLASS MANUFACTURING

Copies of database tables showing all records for glass manufacturing controls,
highlighting revisions.

C-l


-------
Table C-l. CMDB Table 01 Summary

cmname

cmabbreviation

pechanm
eascode

major
poll

controltechnologv

sourcegroup

Sector

Class

equiplile

neidevic
ecode

daterevi
ewed

datasource

Month

s

Description

Cullet Preheat; Glass
Manufacturing—Container

NCLPTGMCN

N0302

NOx

Cullet Preheat

Glass Manufacturing—
Container

ptnonipm

Emerging

10



2013

72 175 182
GM-1





Cullet Preheat; Glass
Manufacturing—Pressed

NCUPHGMPD

N0322

NOx

Cullet Preheat

Glass Manufacturing—
Pressed

ptnonipm

Emerging

10



2013

72 175 182
GM-1





OXY-Firmg; Glass
Manufacturing—General

NDOXYFGMG

N/A

NOx

OXY-Firing

Glass Manufacturing—
General

ptnonipm

Emerging

10





167





Electric Boost; Glass
Manufacturing—General

NELBOGMGN

N0301

NOx

Electric Boost

Glass Manufacturing—
Container

ptnonipm

Known

10



2013

GM-1





Electric Boost; Glass
Manufacturing—Container

NELBOGMCN

N0301

NOx

Electric Boost

Glass Manufacturing—
Container

ptnonipm

Known

10



2006

72 175 182





Electric Boost; Glass
Manufacturing—Flat

NELBOGMFT

N0311

NOx

Electric Boost

Glass Manufacturing—
Flat

ptnonipm

Known

10



2006

72 175 182





Electric Boost; Glass
Manufacturing—Pressed

NELBOGMPD

N0321

NOx

Electric Boost

Glass Manufacturing—
Pressed

ptnonipm

Known

10



2006

72 175 182





Low NOx Burner; Glass
Manufacturing—Container

NLNBUGMCN

N0303

NOx

Low NOx Burner

Glass Manufacturing—
Container

ptnonipm

Known

10

204 205

2013

72 175 179
182 GM-2





Low NOx Burner; Glass
Manufacturing—Flat

NLNBUGMFT

N0312

NOx

Low NOx Burner

Glass Manufacturing—
Flat

ptnonipm

Known

10

204 205

2013

72 175 179
182 GM-2





Low NOx Burner; Glass
Manufacturing—Pressed

NLNBUGMPD

N0323

NOx

Low NOx Burner

Glass Manufacturing—
Pressed

ptnonipm

Known

10

204 205

2006

175 179 18
2





OXY-Firmg; Glass
Manufacturing—General

NOXYFGMGN

N0306

NOx

OXY-Firing

Glass Manufacturing—
Container

ptnonipm

Known

10



2013

GM-1





OXY-Firmg; Glass
Manufacturing—Container

NOXYFGMCN

N0306

NOx

OXY-Firing

Glass Manufacturing—
Container

ptnonipm

Known

10



2006

72





OXY-Firmg; Glass
Manufacturing—Flat

NOXYFGMFT

N0315

NOx

OXY-Firing

Glass Manufacturing—
Flat

ptnonipm

Known

10



2006

72





OXY-Firmg; Glass
Manufacturing—Pressed

NOXYFGMPD

N0326

NOx

OXY-Firing

Glass Manufacturing—
Pressed

ptnonipm

Known

10



2006

72





Selective Catalytic Reduction;
Glass Manufacturing—Container

NSCRGMCN

N03403

NOx

Selective Catalytic
Reduction

Glass Manufacturing—
Container

ptnonipm

Known

10

139

2013

72 172 175
179 182 22
4 GM-2





Selective Catalytic Reduction;
Glass Manufacturing—Flat

NSCRGMFT

N0314

NOx

Selective Catalytic
Reduction

Glass Manufacturing—
Flat

ptnonipm

Known

10

139

2013

72 172 175
179 182 18
6 224 GM-2





Selective Catalytic Reduction;
Glass Manufacturing—Pressed

NSCRGMPD

N0325

NOx

Selective Catalytic
Reduction

Glass Manufacturing—
Pressed

ptnonipm

Known

10

139

2006

72 172 175
179 182 18
6 224





Catalytic Ceramic Filter; Glass
Manufacturing—Flat

CAT



NOx

Catalytic Ceramic
Filter

Glass Manufacturing—
Flat

ptnonipm

Known

20



2013

GM-3






-------
Table C-l. CMDB Table 01 Summary—Description Field

cmabbreviation

description

NCLPTGMCN

Application: This control is the use of cullet preheat technologies to reduce NOx emissions from glass manufacturing operations.
This control is applicable to container glass manufacturing operations classified under 305010402.

NCUPHGMPD

Application: This control is the use of cullet preheat technologies to reduce NOx emissions from glass manufacturing operations.
This control is applicable to pressed glass manufacturing operations classified under 305010404.

NDOXYFGMG

Application: This control is the use of OXY-firing in glass manufacturing furnaces to reduce NOx emissions. Oxygen enrichment refers to the substitution of oxygen for nitrogen in the combustion air used
to burn the fuel in a glass furnace. Oxygen enrichment above 90 percent is sometimes called "oxy-firing."

Discussion: The basic rationale for oxy-firing is improved efficiency, i.e., more of the theoretical heat of combustion is transferred to the glass melt and is not lost in the flue gas. Many other combustion
modification techniques (e.g., flue gas recirculation, staged combustion, and low excess air combustion) reduce NOx formation but also reduce the combustion efficiency. Oxy-firing was originally
developed to improve the combustion efficiency primarily by eliminating the sensible heat lost in heating the nitrogen present in air, which is then lost in the flue gas.

NELBOGMGN

Application: This control is the use of electric boost technologies to reduce NOx emissions from glass manufacturing operations.
This control applies to general glass manufacturing operations classified under SCC 30501401.

NELBOGMCN

Application: This control is the use of electric boost technologies to reduce NOx emissions from glass manufacturing operations.

This control applies to container glass manufacturing operations classified under SCC 30501402.

Discussion: The 250 tons per day plant is assumed to be representative of container glass plants (Pechan, 1998).

NELBOGMFT

Application: This control is the use of electric boost technologies to reduce NOx emissions from glass manufacturing operations.

This control applies to flat glass manufacturing operations classified under SCC 30501403.

Discussion: The 500 tons per day plant is assumed to be representative of flat glass plants (Pechan, 1998).

NELBOGMPD

Application: This control is the use of electric boost technologies to reduce NOx emissions from glass manufacturing operations.
This control applies to pressed glass manufacturing operations classified under SCC 30501403.

Discussion: The 50 tons per day plant is assumed to be representative of pressed glass plants (Pechan, 1998).

NLNBUGMCN

Application: This control is the use of low NOx burner (LNB) technology to reduce NOx emissions. LNBs reduce the amount of NOx created from reaction between fuel nitrogen and oxygen by lowering
the temperature of one combustion zone and reducing the amount of oxygen available in another.

This control is applicable to container glass manufacturing operations classified under 305010402 with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: The 250 tons per day plant is assumed to be representative of container glass plants (Pechan, 1998).

LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply excess
air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air, which acts as a heat sink to lower combustion temperatures (EPA, 2002).

NLNBUGMFT

Application: This control is the use of low NOx burner (LNB) technology to reduce NOx emissions. LNBs reduce the amount of NOx created from reaction between fuel nitrogen and oxygen by lowering
the temperature of one combustion zone and reducing the amount of oxygen available in another.

This control is applicable to flat glass manufacturing operations classified under 305010404 with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: The 500 tons per day plant is assumed to be representative of flat glass plants (Pechan, 1998).

LNBs are designed to "stage" combustion so that two combustion zones are created, one fuel-rich combustion and one at a lower temperature. Staging techniques are usually used by LNB to supply excess
air to cool the combustion process or to reduce available oxygen in the flame zone. Staged-air LNBs create a fuel-rich reducing primary combustion zone and a fuel-lean secondary combustion zone.
Staged-fuel LNBs create a lean combustion zone that is relatively cool due to the presence of excess air, which acts as a heat sink to lower combustion temperatures (EPA, 2002).

(continued)


-------
Table C-l. CMDB Table 01 Summary—Description Field (continued)

cmabbreviation

description

NLNBUGMPD

Application: This control is the use of low NOx burner (LNB) technology to reduce NOx emissions. LNBs reduce the amount of NOx created from reaction between fuel nitrogen and oxygen by lowering
the temperature of one combustion zone and reducing the amo

NOXYFGMGN

Application: This control is the use of OXY-firing in flat glass manufacturing furnaces to reduce NOx emissions. Oxygen enrichment refers to the substituti' 11 ¦ - ygen for nitrogen in the combustion air
used to burn the fuel in a glass furnace. Oxygen enrichment above 90 percent is sometimes called "oxy-firing."

This control applies to general manufacturing operations. This control applies to general glass manufacturing operations classified under SCC 3u5(Jl4(Jl.

Discussion: The basic rationale for oxy-firing is improved efficiency, i.e., more of the theoretical heat of combustion is transferred to the glass melt and is not lost in the flue gas. Many other combustion
modification techniques (e.g., flue gas recirculation, staged combustion, and low excess air combustion) reduce NOx formation but also reduce the combustion efficiency. Oxy-firing was originally
developed to improve the combustion efficiency primarily by eliminating the sensible heat lost in heating the nitrogen present in air, which is then lost m the flue gas.

NOXYFGMCN

Application: This control is the use of OXY-firing in container glass manufacturing furnaces to reduce NOx emissions. Oxygen enrichment refers to the substitution of oxygen for nitrogen in the
combustion air used to burn the fuel in a glass furnace. Oxygen enrichment above 90 percent is sometimes called "oxy-firing."

Discussion: The basic rationale for oxy-firing is improved efficiency, i.e., more of the theoretical heat of combustion is transferred to the glass melt and is not lost in the flue gas. Many other combustion
modification techniques (e.g., flue gas recirculation, staged combustion, and low excess air combustion) reduce NOx formation but also reduce the combustion efficiency. Oxy-firing was originally
developed to improve the combustion efficiency primarily by eliminating the sensible heat lost in heating the nitrogen present in air, which is then lost in the flue gas.

NOXYFGMFT

Application: This control is the use of OXY-firing in flat glass manufacturing furnaces to reduce NOx emissions. Oxygen enrichment refers to the substitution of oxygen for nitrogen in the combustion air
used to burn the fuel in a glass furnace. Oxygen enrichment above 90 percent is sometimes called "oxy-firing."

This control applies to flat-glass manufacturing operations with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: The basic rationale for oxy-firing is improved efficiency, i.e., more of the theoretical heat of combustion is transferred to the glass melt and is not lost in the flue gas. Many other combustion
modification techniques (e.g., flue gas recirculation, staged combustion, and low excess air combustion) reduce NOx formation but also reduce the combustion efficiency. Oxy-firing was originally
developed to improve the combustion efficiency primarily by eliminating the sensible heat lost in heating the nitrogen present in air, which is then lost in the flue gas.

NOXYFGMPD

Application: This control is the use of OXY-firing in pressed glass manufacturing furnaces to reduce NOx emissions. Oxygen enrichment refers to the substitution of oxygen for nitrogen in the combustion
air used to burn the fuel in a glass furnace. Oxygen enrichment above 90 percent is sometimes called "oxy-firing."

Discussion: The basic rationale for oxy-firing is improved efficiency, i.e., more of the theoretical heat of combustion is transferred to the glass melt and is not lost in the flue gas. Many other combustion
modification techniques (e.g., flue gas recirculation, staged combustion, and low excess air combustion) reduce NOx formation but also reduce the combustion efficiency. Oxy-firing was originally
developed to improve the combustion efficiency primarily by eliminating the sensible heat lost in heating the nitrogen present in air, which is then lost in the flue gas.

(continued)


-------
Table C-l. CMDB Table 01 Summary—Description Field (continued)

cmabbreviation

description

NSCRGMCN

Application: This control is the selective catalytic reduction of NOx through add-on controls. SCR controls are post-combustion control > i n i> ¦¦1 !>.¦¦¦ i1 ised on the chemical reduction of nitrogen oxides
(NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal efficiency, which aiiows ine process to occur at lower temperatures.

Applies to glass-container manufacturing processes, classified under SCC 30501402 and uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is typically implemented
on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the rate of reaction (EPA,
2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a specific temperature range and in the presence
of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the decrease in reaction
temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety issues with the use of
anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support mi 11< mi ¦ Most catalyst formulations contain additional compounds or sup-ports, providing thermal
and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction temperature range; residence
time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx concentration level; molar ratio of injected reagent to
uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst deactivation; and catalyst management (EPA, 2001).

NSCRGMFT

Application: This control is the selective catalytic reduction of NOx through add-on controls. SCR controls are post-combustion control technologies based on the chemical reduction of nitrogen oxides
(NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal efficiency, which allows the process to occur at lower temperatures.

Applies to large(>l ton NOx per OSD) flat-glass manufacturing operations (SCC 30501403) with uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is typically implemented
on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the rate of reaction (EPA,
2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a specific temperature range and in the presence
of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR. the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the decrease in reaction
temperature and increase in efficiency is accompanied by a significant increase m capital and operating costs (EPA, 2002). The cost increase is due to the large amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety issues with the use of
anhydrous ammonia, as it must be transported and stored under pressure ("EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support structure. Most catalyst formulations contain additional compounds or sup-ports, providing thermal
and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction temperature range; residence
time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx concentration level; molar ratio of injected reagent to
uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst deactivation; and catalyst management (EPA, 2001).

(continued)


-------
Table C-l. CMDB Table 01 Summary—Description Field (continued)

cmabbreviation

description

NSCRGMPD

Application: This control is the selective catalytic reduction of NOx through add-on controls. SCR controls are post-combustion control > i n i> ¦¦1 !>.¦¦¦ i1 ised on the chemical reduction of nitrogen oxides
(NOx) into molecular nitrogen (N2) and water vapor (H20). The SCR utilizes a catalyst to increase the NOx removal efficiency, which aiiows ine process to occur at lower temperatures.

Applies to pressed-glass manufacturing operations, classified under SCC 30101404 and uncontrolled NOx emissions greater than 10 tons per year.

Discussion: Selective Catalytic Reduction (SCR) has been widely applied to stationary source, fossil fuel-fired, combustion units for emission control since the early 1970s. SCR is typically implemented
on units requiring a higher level of NOx control than achievable by SNCR or other combustion controls (EPA, 2002).

Like SNCR, SCR is based on the chemical reduction of the NOx molecule. The primary difference between SNCR and SCR is that SCR uses a metal-based catalyst to increase the rate of reaction (EPA,
2002). A nitrogen based reducing reagent, such as ammonia or urea, is injected into the flue gas. The reagent reacts selectively with the flue gas NOx within a specific temperature range and in the presence
of the catalyst and oxygen to reduce the NOx.

The use of a catalyst results in two advantages of the SCR process over SNCR, the higher NOx reduction efficiency and the lower and broader temperature ranges. However, the decrease in reaction
temperature and increase in efficiency is accompanied by a significant increase in capital and operating costs (EPA, 2002). The cost increase is due to the large amount of catalyst required.

The SCR system can utilize either aqueous or anhydrous ammonia as the reagent. Anhydrous ammonia is a gas at atmospheric pressure and normal temperatures. There are safety issues with the use of
anhydrous ammonia, as it must be transported and stored under pressure (EPA, 2002). Aqueous ammonia is generally transported and stored at a concentration of 29.4% ammonia in water.

Today, catalyst formulations include single component, multi-component, or active phase with a support mi 11< uk Most catalyst formulations contain additional compounds or sup-ports, providing thermal
and structural stability or to increase surface area (EPA, 2002).

The rate of reaction determines the amount of NOx removed from the flue gas. The important design and operational factors that affect the rate of reduction include: reaction temperature range; residence
time available in the optimum temperature range; degree of mixing between the injected reagent and the combustion gases; uncontrolled NOx concentration level; molar ratio of injected reagent to
uncontrolled NOx; ammonia slip; catalyst activity; catalyst selectivity; pressure drop across the catalyst; catalyst pitch; catalyst deactivation; and catalyst management (EPA, 2001).

CATCFGMFT

Application: Filter tubes have nanobits of proprietary catalyst are embedded throughout the filter walls. The system can achieve excellent NOx removal using liquid ammonia that is injected upstream of the

filters, reacting with NOx at the catalyst to form nitrogen gas and water vapor.

This control applies to general glass manufacturing operations classii 10		 i> r SCC 30501403


-------
Table C-2. CMDB Table 02 Efficiencies

cmabbreviatio
n

polluta
nt

loca
le

Effec
tive
Date

existing
measure
abbr

neiexistingd
evcode

minemissi
ons

maxemissi
ons

controleffi
ciency

costyea
r

costperton

ruleeff

rulepen

equation
type

capii rl'.iri
or

(llM-Hll

ntrate

r.ipannr

atio

incremen
talcpt

details

NCLPTGMCN

NOx







0

365

0

5

2002

5000

100

100

cpton

0.1424



4.5



Applied to large source types

NCLPTGMCN

NOx







0

0

365

5

2002

5000

100

100

cpton

0.1424



4.5



Applied to small source types

NCUPHGMPD

NOx







0

365



5

2002

5000

100

100

cpton

0.1424



4.5



Applied to large source types

NCUPHGMPD

NOx







0

0

365

5

2002

5000

100

100

cpton

0.1424



4.5



Applied to small source types

NELBOGMCN

NOx







0

365



10

1990

7150

100

100

cpton

0.1424



0



Applied to large source types

NELBOGMCN

NOx







0

0

365

10

1990

7150

100

100

cpton

0.1424



0



Applied to small source types

NELBOGMFT

NOx







0

365



10

1990

2320

100

100

cpton

0.1424



0



Applied to large source types

NELBOGMFT

NOx







0

0

365

10

1990

2320

100

100

cpton

0.1424



0



Applied to small source types

NELBOGMPD

NOx







0

365



10

1990

8760

100

100

cpton

0.1424



0



Applied to large source types

NELBOGMPD

NOx







0

0

365

10

1990

2320

100

100

cpton

0.1424



0

8760

Applied to small source types

NELBOGMGN









0

365

0

30

2002

7100

100

100

cpton

0.1424



0



Applied to large source types

NELBOGMGN









0

0

365

30

2002

7100

100

100

cpton

0.1424



0



Applied to small source types

NLNBUGMCN

NOx







0

365



40

2007

1072

100

100

cpton

0.14



4.3

1690

Applied to large source types

NLNBUGMCN

NOx







0

0

365

40

2007

1365

100

100

cpton

0.14



4.2

1690

Applied to small source types

NLNBUGMFT

NOx







0

0

365

40

2007

574

100

100

cpton

0.14



4.2



Applied to small source types

NLNBUGMFT

NOx







0

365



40

2007

447

100

100

cpton

0.14



4.3



Applied to large source types

NLNBUGMPD

NOx







0

365



40

1990

1500

100

100

cpton

0.1424



2.2



Applied to large source types

NLNBUGMPD

NOx







0

0

365

40

1990

1500

100

100

cpton

0.1424



2.2



Applied to small source types

NOxYFGMCN

NOx







0

0

365

85

1990

4590

100

100

cpton

0.1424



2.7



Applied to small source types

NOxYFGMCN

NOx







0

365



85

1990

4590

100

100

cpton

0.1424



2.7



Applied to large source types

NOxYFGMFT

NOx







0

365



85

1990

1900

100

100

cpton

0.1424



2.7



Applied to large source types

NOxYFGMFT

NOx







0

0

365

85

1990

1900

100

100

cpton

0.1424



2.7



Applied to small source types

NDOXYFGMG

NOx







0





85

1999

4277

100

100

cpton











NOxYFGMPD

NOx







0

0

365

85

1990

3900

100

100

cpton

0.1424



2.7



Applied to small source types

NOxYFGMPD

NOx









365



85

1990

3900

100

100

cpton

0.1424



2.7



Applied to large source types

NOxYFGMGN











365

0

85

2002

2353

100

100

cpton

0.1424



2.7



Applied to large source types

NOxYFGMGN











0

365

85

2002

2353

100

100

cpton

0.1424



2.7



Applied to small source types

NSCRGMCN

NOx







0

365

0

75

2007

1684

100

100

cpton

0.1424



4.2



Applied to large source types

NSCRGMCN

NOx







0

0

365

75

2007

2169

100

100

cpton

0.1424



4.5



Applied to small source types

NSCRGMFT

NOx







0

365

0

75

2007

855

100

100

cpton

0.1424



3.7

710

Applied to large source types

NSCRGMFT

NOx







0

0

365

75

2007

957

100

100

cpton

0.1424



3.4



Applied to small source types

(continued)


-------
Table C-2. CMDB Table 02 Efficiencies (continued)

cmabbreviatio
n

polluta
nt

loca
le

Effec
tive
Date

existing
measure
abbr

neiexistingd
evcode

minemissi
ons

maxemissi
ons

controleffi
ciency

costyea
r

costperton

ruleeff

rulepen

equation
type

capii rl'.iri
or

(llM-Hll

ntrate

r.ipannr

atio

incremen
talcpt

details

NSCRGMPD

NOx







0

365



75

1990

2530

100

100

cpton

0.1424



1.3



Applied to large source types

NSCRGMPD

NOx







0

0

365

75

1990

2530

100

100

cpton

0.1424



1.3



Applied to small source types

CATCFGMFT

NOx







0

365

0

95

2013

997

100

100

cpton

0.05



4.6



Applied to large source types

CATCFGMFT

NOx







0

0

365

95

2013

1045

100

100

cpton

0.05



4.6



Applied to small source types

O


-------
Table C-3. CMDB Table 06 References (New)

Data Source

Description

GM-1

Oxygen Enriched Air Staging a Cost-effective Method For Reducing NOx Emissions. Industrial Tcchnologics. April 2002. Available at:
htto://wwwl.eere.enerev.eov/manufacturine/resources/elass/odfs/airstaeine.i3df

GM-2

Best Available Techniques (BAT) Reference Document for the Manufacture of Glass. European Commission 2013. Available at:
htto://eiirocb.irc.ec.eurora.eu/reference/BREF/GLS Adootcd 03 2012.odf

GM-3

Confidential Vendor Quote

Table C-4. CMDB Table 04_Equationsa

cmabbreviation

cmeqntype

pollutant

costyear

varl

var2

var3

var4

var5

var6

var7

var8

var9

varlO

NLNBUGMCN

Type 2

NOx

2008

30,930

0.45

9,377

0.40













NLNBUGMFT

Type "L"

NOx

2008

527

664,557

132

150,105













NSCRGMCN

Type 2

NOx

2008

79,415

0.51

















NSCRGMCN

Type "L"

NOx

2008





643

135,302













NSCRGMFT

Type "L"

NOx

2008

3,681

1,000,000

842

424,930













aType "L" is a linear equation; variables are the slope and intercept.


-------
APPENDIX D
LEAN BURN ENGINES

Copies of the database tables for showing all records for Lean Burn Engine NOx controls
are provided:

-	Table D-01_Summary

-	Table D-02_Efficiencies

-	Table D-03_SCCs

-	Table D-04_Equations

-	Table D-06 References

D-l


-------
Table D-01_Summary

cmname

cmabbreviation

pechanme
ascode

majorp
oil

controltechn
ology

sourcegr
oup

sector

class

equiplife

neid
evic
eco
de

datereviewe
d

datasour
ce

months

description

Low

Emission
Combustion;
Lean Burn

ICE—NG

NLECICENG



NOx

Low

Emission
Combustion

Lean
Burn

ICE—
NG

PTNONIPM

Known

10



9/15/2013

ABCD3



Low Emission Combustion includes Precombustion
chamber head and related equipment on a Lean Burn
engine.

Layered
Combustion;
Lean Burn

ICE 2

stroke—NG

NLCICE2SNG



NOx

Layered
Combustion

Lean
Burn

ICE—
NG

PTNONIPM

Known

10



9/15/2013

ABCD1



Layered combustion—2 stroke, Lean Burn, NG (Air
Supply; Fuel Supply; Ignition; Electronic Controls;
Engine Monitoring). Evaluation for 3 most representative
made/models of 2 stroke LB compressor engines. All
retrofit combustion-related controls may not be available
for all manufacturers and models of 2-stroke lean burn
engines. Actual NOx emission rates would be engine
design specific. Efficiency achieved may range from 60
to 90%, depending on the make/model of engine
(approximate range of NOx emissions of 3.0 to 0.5
g/bhp-hr).

Layered
Combustion;
Lean Burn
ICE 2 stroke
Large
Bore—NG

NLCICE2SLBNG



NOx

Layered
Combustion

Lean
Burn

ICE—
NG

PTNONIPM

Known

10



9/15/2013

ABCD1



Layered combustion—for Large Bore, 2 stroke, Lean
Burn, Slow Speed (High Pressure Fuel Injection achieves
90% reduction; Turbocharging achieves 75% reduction;
Precombustion chambers achieves 90% reduction;
Cylinder Head Modifications). All retrofit combustion-
related controls may not be available for all
manufacturers and models of 2-stroke lean burn engines.
Actual NOx emission rates would be engine design
specific. Efficiency achieved may range from 60 to 90%,
depending on the make/model of engine (approximate
range of NOx emissions of 3.0 to 0.5 g/bhp-hr).

Air to Fuel
Ratio

Controller;
Lean Burn

ICE—NG

NAFRCICENG



NOx

Air to Fuel

Ratio

Controller

jn Li-

1 NIPM

Known

10



12/5/2012

ABCD3





Selective
Catalytic
Reduction;
Lean Burn
ICE 4

Stroke—NG

NSCRICE4SNG



NOx

Selective
Catalytic
Reduction

Lean
Burn

ICE—
NG

PTNONIPM

Known

10



9/15/2013

ABCD1
ABCD2
ABCD3



SCR can be used on Lean Burn, NG engines. Assumed
SCR can meet NOx emissions of 0.89 g/bh-hr. This is a
Known technology, however there is indication that
applicability is engine/unit specific.

Selective
Catalytic
Reduction;
ICE—Diesel

NSCRICEDS





. ".'i<;ctive
Catalytic
Reduction

ICE—
Diesel

PTNONIPM

Known

7



9/15/2013

ABCD4



SCR can be used on Diesel engines.


-------
Table D-02 Efficiencies

cmabbreviation

polluta
nt

local
e

Effec
tlve
Date

existing
measur
eabbr

neiexi
stingd
evcod
e

mine
missio

IIS

maxemis
sions

controle
fflciency

costyear

costperton

raleeff

rulepe
n

equation
type

caprect'ac
tor

discount

rate

capann
ratio

increme
ntalcpt

details

NLECICENG

NOx

NA

NA

NA

NA

0

365

80

2001

1,000

100

100

cpton

0.1424

7

7.025

NA



NLCICE2SNG

NOx

NA

NA

NA

NA

0

365

97

2009

4,900

100

100

cpton

0.1424

7

7.024

NA



NLCICE2SLBNG

NOx

NA

NA

NA

NA

365

0

97

2010

1,500

100

100

cpton

0.1424

7

7.024

NA

Apply to large
source types.
Assumed Interest
Rate of 7 percent
(not provided in
documentation)
to calculate
annual costs.

NLCICE2SLBNG

NOx

NA

NA

NA

NA

0

365

97

2010

38,000

100

100

cpton

0.1424

7

7.024

NA

Apply to small
source types.

NAFRCICENG

NOx

NA

NA

NA

NA

0

365

80

2001

200

100

100

cpton

0.1424

7

7.023

NA



NSCRICE4SNG

NOx

NA

NA

NA

NA

0

365

96

2001

2,900

100

100

cpton

0.1424

7

1.401

NA



NSCRICEDS

NOx

NA

NA

NA

NA

0

365

90

2005

9,300

100

100

cpton

0.1098

7

2.45

NA




-------
Table D-03 SCCs

cmabbreviation

Source Oassiflcation Code

Status

NLECICENG

20200252



NLECICENG

20200254



NLECICENG

20200255



NLECICENG

20200256



NLCICE2SNG

20200252



NLCICE2SNG

20200254



NLCICE2SNG

20200255



NLCICE2SNG

20200256



NLCICE2SLBNG

20200252



NLCICE2SLBNG

20200254



NLCICE2SLBNG

20200255



NLCICE2SLBNG

20200256



NLCICE2SLBNG

20200401



NLCICE2SLBNG

20200402



NLCICE2SLBNG

20200403



NAFRCICENG

20200252



NAFRCICENG

20200254



NAFRCICENG

20200255



NAFRCICENG

20200256



NSCRICE4SNG

20200252



NSCRICE4SNG

20200254



NSCRICE4SNG

20200255



NSCRICE4SNG

20200256



NSCRICEDS

20200102



NSCRICEDS

20200107




-------
Table D-04_Equations

cmabbreviation

cmeqntype

pollutant

co sty ear

varl

var2

var3

var4

var5

var6

var7

var8

var9

varlO

NSCRICE4SNG

linear capital and annual

NOx

2001

107.1

27186

83.64

14718













NLECICENG

capital and annual

NOx

2001

16019

0.0016

2280.8

0.0016













NAFRCICENG

linear capital and annual

NOx

2001

1.0337

4354.5

0.1852

619.99













Table D-06 References

Data Source

Description

ABCD1

OTC 2012. Technical Information Oil and Gas Sector. Significant Stationary Sources of NOx Emissions. Final. October 17, 2012.

ABCD2

SJVAPCD 2003. RULE 4702—Internal Combustion Engines—Phase 2. Appendix B, Cost Effectiveness Analysis for Rule 4702
(Internal Combustion Engines—Phase 2). San Joaquin Valley Air Pollution Control District. July 17, 2003.
www.arb.ca.sov/om/i3mmeasures/ceffect/rulcs/sivaDcd 4702.Ddf

ABCD3

CARB 2001. Determination of Reasonably Available Control Technology and Best Available Retrofit Control Technology for
Stationary Spark-Ignited Internal Combustion Engines. California Environmental Protection Agency, Air Resources Board, Stationary
Source Division, Emissions Assessment Branch. Process Evaluation Section. November 2001.

ABCD4

EPA 2010. Alternative Control Techniques Document: Stationary Diesel Engines. March 5, 2010.


-------
APPENDIX E

NOTES PROVIDED HERE TO EPA QUESTIONS ON LEAN BURN RICE

E-l


-------
EPA Question 1: What is the applicability of SCR to RICE, especially Lean Burn?

Notes for Question 1

In addition to the two documents cited in Section 5 of the report with costs for selective
catalytic reduction (SCR) for Lean Burn (LB) engines, there are several other references that
indicate SCR is feasible for LB engines and several that provide input on technical issues related
to SCR use for LB engines. In summary, from the references reviewed, SCR seems to be
technically feasible in most instances for LB engines, however, SCR application may not be
feasible in all cases due to technical issues at individual sites and individual engines. In addition,
SCR costs are higher relative to other NOx control techniques for LB engines. See more detailed
discussion below.

SCR can be applied to LB engines, achieving greater than 90 percent NOx reductions
(Table 4 on p. 6 provides a slightly different value, greater than 95 percent). The costs [assumed
this referred to capital costs] ranged from $50/hp to $ 125/hp. No annual operating costs were
provided. In discussions on p. 8 regarding "catalysts on 1C engines" in general (including NSCR,
SCR, oxidation, and Lean-NOx), it is noted that "Thousands of stationary 1C engine catalyst
applications have been effectively used for stationary 1C engine gaseous emission control for
five years or more. Some installations, however, do experience performance loss over time,"
however the text goes on to explain remedies for catalyst poisoning issues. Costs [capital] for
SCR, LB ranged from $50 to $ 125/hp (no cost year provided). (MECA 1997)

The literature suggests that SCR is technically feasible for LB engines but there are
problems that make SCR installation questionable. Two stroke (2S) LB engines are sensitive to
changes in exhaust pressure, which could be problematic for retrofit of SCR on existing engines,
but can be alleviated with proper design and sizing of airflow and exhaust components. This
reference cited a presentation that indicated the following issues with SCR: applying SCR to
pipeline engines is not feasible because the exhaust temperatures (T) are below the operating
window for SCR or where SCR effectiveness is reduced; SCR installations are at unmanned
facilities; and SCR has not been demonstrated for variable loads. However, the OTC 2012
reference responded to each of these issues, stating that there are several manufacturers and
suppliers that offer SCR systems that indicate their catalysts are capable of effectively operating
over a wide range of exhaust gas T; modern software based controls and SCADA
communication technologies allow operation from a remote location; and SCR can function
properly over a broad range of loads given catalysts that are effective over wide T ranges,
modern controls regulate fuel and air flows to ensure combustion O2 and T are at expected levels

E-2


-------
and to regulate reagent flow. A study conducted for retrofitting existing pipeline engines
indicates that SCR is a high cost alternative to combustion improvements, primarily due to the
high cost of ongoing reagent consumption, (p. 25-26) (There is a similar discussion for SCR for
four stroke (4S) LB on p. 39-40; cited presentation at Gas Machinery Conference in October
2011.) (OTC 2012)

Shell indicated they have installed SCR on diesel engines (LB) that they utilize in drilling
rig operations. Shell indicated that have been able to achieve greater than 90 percent reduction in
NOx emissions while encountering minimal operational issues (see p. 10). (OTC 2012)

The OTC 2012 document indicated that MECA has noted there have been limited
examples to date of SCR retrofit on 2S LB engines as demonstration test programs, but the
results of these programs have not been published (see p. 27). It appears that SCR for NOx does
not appear to be technically infeasible genetically but that individual 2S LB engine
characteristics and installations may be greatly problematic or not cost effective, although this
site-specific issue is not altogether different than other emission reduction technologies (see p.27,
40). (OTC 2012)

The OTC 2012 document indicated that MECA has stated the commercial use of SCR
systems for LB stationary engines have been in place since the mid-1980's in Europe and since
the early 1990s in the US. One MECA member company has installed over 400 SCR systems
worldwide for stationary engines with varying fuel combinations, including dozens of NG
compressor engines in the US. These 4S LB engines with urea-SCR achieve >90% NOx
reduction (see p.40). (OTC 2012)

EF&EE announced in November 2010 that is received an order from Clean Air Power
Inc. for 6 SCR systems, to be installed on large LB NG compressor engines at gas storage sites in
TX and MS (see p.40). (OTC 2012)

Clean Air Power cited: 4 SCRs supplied at Pine Prairie Energy Center, Louisiana; 1 SCR
supplied at EXTERRAN/TRESPALACIOS, Texas; and 4 SCRs supplied to EXTERRAN/LEAF
River, Mississippi (see p.41). (OTC 2012)

A PowerPoint slide presentation from a MARAMA workshop discusses the use of SCR
for RICE and LB. Johnson Mathey (JM) included SCR as a feasible control for LB engines in a
presentation at a May 2011 MARAMA Workshop. (The SCR systems included Urea and
Ethanol as reagents.) SCR operating temperatures range from 700 to 900°F for internal
combustion (IC) engines and achieved 90 percent NOx reductions. The budgetary costs

E-3


-------
[assumed this referred to capital costs] ranged from $150/ hp for a 500 hp unit (approximately
$75,000) to $42/hp for a 3000 hp unit (approximately $126,000) (cost year not provided). No
annual operating costs were provided. JM cited 4 LB engine installations of SCR on gas
compressors at 2 locations, including Loudon Compressor Station in Clarksburg, WV and Lodi
Compressor/Storage in Kirby Hills, CA. (Chu 2011) These engines are listed in the following
table:

SCR for Lean Burn Engines—Johnson Mathey presentation at 2007 MAUA.M A Workshop

Engine Model

Engine hp

NOx, g/bhp-hr

NOx Reduction, %

CAT G3516

1,340

1.5

90%

CAT G3608

2,370

0.7

90%

CAT G3612

3,550

0 "

90%

CAT G3616

4,735

0 "

90%

References

(MECA 1997). Emission Control Technology for Stationary Internal Combustion Engines:
Status Report. Manufacturers of Emission Controls Association (MECA). July 1997.

(Chu 2011). NOx Control for Stationary Gas Engines. W. Chu, Johnson Mathey. Presented at
Advances in Air Pollution Control Technology, MARAMA Workshop. May 19, 2011.

(OTC 2012). Technical Information Oil and (}as Sector, Significant Stationary Sources of NOx
Emissions. Final. October 17, 2012.

E-4


-------
EPA Question 2: What are credible estimates of the percentage of RICE NOx Emissions
that are lean burn versus rich burn when RICE emissions are unspecified?

Notes for Question 2

There does not seem to be much information on NOx emission totals for LB and rich
burn (RB) engines. A few references attempted to provide information on the numbers or
populations of LB and RB engines. Several of the references highlighted surveys of engine
populations and summary information from various engine databases. These data in general tend
to point to a large LB engine population, however most of the references noted that RB engines
are typically not captured or covered in surveys, databases, or by permits because the RB engines
tend to be smaller in size. In general, larger engines tend to be LB and smaller engines tend to be
RB. The ERLE 2009 study noted that approximately 73% of the 5,600 engines/horsepower
capacity covered in their study of NG pipeline systems are LB, and approximately 6% are RB
(the balance is not known). In the KSU 2011 database, approximately 66% of the 4,729 engines
used in E&P at major sources are LB and 34% are RB. In addition, the EDF 2008 document
cited a 2007 survey conducted for DFW NAA and AA that attempted to identify those engines
that did not meet reporting requirement thresholds and were therefore not included in the TCEQ
inventory. This reference, which included small engines, indicated that for smaller engines <500
hp, approximately 96% are RB and 4% are LB. The reference also indicated that for larger
engines >500 hp, there is approximately a 50-50 split of LB and RB engines and of horsepower
capacity. The ETCG 2013 reference also highlights engines in the Barnett Shale region. Data
from the TCEQ Barnett Shale Special Inventory (Phase I) survey indicated that the majority of
engines in the Barnett Shale are RB (84%). For those engines <240 hp, 95% are RB and 5% are
LB, however, in looking at those engines >240 hp, 59% are LB and 41% are RB. More details
for each of these references are provided in the discussion that follows.

Note also that the emissions rate in g/bhp-hr for LB engines tend to be higher, and the
emissions rate for RB engines tends to be lower. (See the tables under Question 4 of this
appendix for relative emission rate values for LB and RB engines in various states and local
districts.)

A summary of the information available from various references is provided below.

The CARB 2001 reference indicated that LB engines tend to be larger in size, and smaller
engines tend to be RB (p.B-4). (CARB 2001)

EPA received comments from the Interstate Natural Gas Association of America
(INGAA) on the 2002 proposed rule, where EPA indicated that 156 of 168 large engines listed in

E-5


-------
the NOx SIP Call Inventory that have SIC codes associated with the NG transmission industry
are LB engines (with the exception that the other 12 engines are no longer in service, are owned
by a company not included in the industry database, or are duplicates). INGAA recommended
that EPA assume all large NG stationary engines in the inventory are LB. (EPA 2003).

One prominent use of large Reciprocating Internal Combustion Engines (RICE) is to
drive NG pipeline compressor stations; almost all engines affected by the NOx SIP Call Phase 2
rule in IL (except for 3 engines) are used to compress NG at NG pipeline stations (11 ¦ l\\ 2< n >7)

A 2009 ERLE study cited in this reference indicated there are 5,600 engines on the NG
pipeline systems with a collective rating of 9,150,000 hp. That study further indicated that
approximately 80 percent of the rated output was low speed 2S, low speed 4S integral engines
and diesel medium speed engines converted to spark ignition (SI). Of these 80 percent of
engines, 78 percent were 2S LB, 14 percent were 4S LB, and 8 percent were 4S RB. (On a rated
horsepower basis, 80 percent was 2S LB, 15 percent was 4S LB, and 5 percent 4S RB) (p. 16).
[On an overall basis, compared to the full 9,150,000 hp collective rating, 2S LB would be
roughly 62%, 4S LB would be roughly 1 1%, and 4S RB would be roughly 6% of the overall
rating/engines. So 73% would be LB, 6% would be RB, and the balance is not known.] (OTC
2012)

Engine Type

No. Engines, %

Horsepower, %

2S LB

78

80

4S LB

14

15

4S RB

8

5

The DE 2012 document cited a 2003 Pipeline Research Council International (PRCI)
document that identified 5,686 engines: 71% are LB and 29% are RB (based on dropping the
turbine numbers in the table below) (p. 19). (DE 2012) [These data may be repeated in OTC
2012, as it looks fairly similar to the 2009 ERLE study data cited above from OTC 2012.]

2003 Pipeline Research Council International Data (PRCI)

Unit Type

U.S. Total Units (%)

Avg hp

2SLB

2,955 (44%)

2,113

4SLB

1,059 (16%)

1,844

RB

1,672 (25%)

589

Turbine

1,016 (15%)

6,121

E-6


-------
Energy Information Agency (EIA) data cited in the OTC 2012 reference indicated there
were 1201 NG mainline compressor stations in the U.S. in 2006, with combined rating of
16,800,000 hp. Between 2007 and 2010, the Federal Energy Regulation Commission (FERC)
approved new compressor stations or upgrades to existing compressor facilities that were
expected to add 2,600,000 hp (p. 16). (OTC 2012)

The Kansas State University (KSU) 2011 document included a database on 4.721) engines
used in Exploration and Production (E&P) at major sources. LB engines accounted for bb
percent of engines (17 percent are 2S and 49 percent are 4S), and RB engines accounted for 34
percent. LB outnumbers RB among engines included in the database; because many engines
rating less than 100 hp are not included, and because the majority of the smaller units are 4S RB,
RB are actually underrepresented in the database. A listing of the engines (manufacturer and
model), air to fuel (A/F) ratio type, cycle, and horsepower are included in Appendix I of the KSU
2011 document. The database was not meant to collect every single engine in use but rather to
provide a frequency distribution of engines. The data was pulled from multiple sources,
including the State of Wyoming Engine Inventory Database, EPA 1CCR Database, GTI/PRCI
Engine and Turbine Database, and Database of Colorado and New Mexico Engines (from
Universal Compression). The engine database likely includes only permitted engines, and lower-
hp engines are underrepresented in the database, (pp. 5-7) (KSU 2011)

The EDF 2008 reference indicated most engines in Barnett Shale area of Texas are 100 to
500 hp but some large engines of 1000+ hp are also used. (EDF 2008)

The EDF 2008 reference indicated that the TCEQ Point Source Emissions Inventory
(PSEI) does not include a substantial fraction of compressor engine emissions. Most of the
missing engines in the DFW NAA were units with emissions below the reporting thresholds, but
the combined emissions from large numbers of these engines can be substantial (pp. 13-14). The
2007 DFW Engine survey indicated there were approximately 680,000 hp of installed engine
capacity in DFW NAA not previously reported to the TCEQ PSEI (p. 14). The report also
estimated that there is approximately 132,000 hp of engines in Attainment Area (AA) counties
within the Barnett Shale that don't report to PSEI (non-PSEI) (p. 14). The LB and RB engine
data from the 2007 DFW Engine Survey for the DFW NAA is provided in the table below. In
this survey, there seem to be fairly even numbers of LB (51%) and RB (49%) engines in the
>500 hp category, and there seems to be fairly even horsepower capacity for the LB and RB
engines. For smaller engines that are <500 hp, there are significantly more RB engines (736

E-7


-------
engines, or 96%) than LB engines (27 engines, or 4%). In addition, for the smaller engines <500
hp, the horsepower capacity for RB represents 15% and for LB is <1%. (EDF 2008)

Installed Engine Capacity in 2007 DFW Engine Survey by Engine Type and Size, in DFW
NAA (EDF 2008)

Engine
Type

Engine
Size, hp

Number of
Engines

Percent of
Engines,

%

Typical Size,
hp

Installed
Capacity, hp

Percent of
Installed
Capacity,

%

RB

<50

12

1.03%

50

585

0.086%

RB

50-500

724

62%

140

101,000

15%

RB

>500

200

17%

1,400

280,000

41%

LB

<500

27

2.3%

185

4.940

0.72%

LB

>500

206

18%

1.425

294.000

44%

The EDF 2008 reference looked at all of the compressor engines in the Barnett Shale
region, including both the engines located within the DFW NAA and the engines in the DFW
AA (including those larger engines that report to the PSE1 and those non-PSEI engines). New
TCEQ rules became effective in 2009 to reduce NOx from the subset of engines located in the
DFW NAA that typically are not reported to the PSE1 (due to their small size) for major sources
(p. 25). Engines that are located outside the DFW NAA are not subject to the 2009 rule. As
shown in the table below, a 50% reduction of emissions from 2007 to 2009 was estimated in
DFW NAA, taking into account the growth, regulation affect, and NSCR installations. For AA
engines, emissions will increase from 2007 to 2009 due to growth and the fact that no regulation
applies (these engines not subject to 2009 engine regulation) (p. 19). (EDF 2008)

NOx Emissions from Compressor Engines in Barnett Shale of Texas (EDF 2008)

Area

2007 NOx Emissions, tpd

2009 NOx Emissions, tpd

DFW NAA engines

32

16

AA engines

20

31

Barnett Shale engines, total

52

47

The reference then looked at emission reductions for extending the 2009 rule to all
engines in the Barnett Shale (including those in the AA). By extending the 2009 engine rule,

E-8


-------
NOx emissions from AA engines would drop by approximately 6.5 tpd (p.25) (this approach
reduces emissions from a large number of engines, in particular RB engines between 50 to 500
hp). (EDF 2008)

The ETCG 2013 reference indicated that analysis of test reports at the TCEQ Tyler office
showed 68 compressor engines: 9 engines (13%) <240 hp and 59 engines (87%) >240 hp (and
69% of all engines >500 hp) (p. 11). (A graph showing the distribution of the hp for all 68
engines is shown on p. 12 of the reference document.) (ETCG 2013)

The ETCG 2013 reference discussed TCEQ Barnett Shale Special Inventory (Phase 1)
survey data. The table below is a summary of the engine horsepower distribution. (A graph
showing the distribution of NG engines in the Barnett Shale region is shown in Figure 5-1 on
p. 21 of the reference document.) The majority of engines in the Barnett Shale are RB and are
<240 hp, see the two tables below. This data set shows that smaller hp engines are predominantly
RB, with 2,089 engines <240 hp are RB (95%) and 104 engines (5%) are LB. For engines >240
hp, 327 engines (59%) are LB and 230 engines (41%) are RB.

2009 Equipment Inventory of Stationary NG Engines by Horsepower for Barnett Shale
Region. (ETCG 2013)

Percent of

Engine Size

Total I'.ngines

Total
Engines

Engine Type,
RB or LB

Number of
Engines

Percent of Each
Size Category

0 to 50 hp

31"

i:".,

RB

302

95%







LB

15

4.7%

50 to 240 hp

I.S"(.

«.x%

RB

1,787

95%







LB

89

4.7%

>240 hp

55"

20%

RB

230

41%







LB

327

59%

E-9


-------
Barnett Shale Special Inventory Phase I Equipment Survey Data on Stationary Gas-Fired
Engines for 2009. (ETCG 2013)

Engine counts
<240 hp	>240 hp	Total

RB and LB

2,193	557	2,750

RB only

2,089	230	2,319

LB only

104	327	431

The CO DPHE reference indicates that large NG RICE represent 16% of the statewide
point source NOx emissions (16,199 tpy of 101,818 tpy) and 73% of the ICENOx emissions
(16,199 tpy of 22,210 tpy) (p. 1). (CO DPHE)

Example Emissions Estimates: It is difficult to draw conclusions for the emissions from
LB versus RB from the data provided. However, some assumptions could be made to help draw
conclusions for the defined scenario. If assume that the total capacity between LB and RB in the
ERLE study is more representative of the total reporting population than the 50-50 split in the
EDF study; assume that operating hours are similarly distributed for both LB and RB; and if the
EFs tend to be higher for LB than for RB engines, then it is likely that 90% plus of the total
emissions are from I ,li

References

(CARB 2001). / V/iT///ii ml ioii of Reasonably Available Control Technology and Best Available
Retrofit (. 'onirol technology for Stationary Spark-Ignited Internal Combustion Engines.
California Environmental Protection Agency, Air Resources Board, Stationary Source
Division, Emissions Assessment Branch, Process Evaluation Section. November 2001.

(IEPA 2007). Technical Support Document for Controlling NOx Emissions from Stationary
Reciprocating Internal Combustion Engines and Turbines. AQPSTR 07-01. Illinois
Environmental Protection Agency, Air Quality Planning Section, Division of Air
Pollution Control, Bureau of Air. March 19, 2007.

(EPA 2003). Stationary Reciprocating Internal Combustion Engines: Technical Support

Document for NOx SIP Call. U.S. Environmental Protection Agency. D. Grano and B.
Neuffer. October 2003.

E-10


-------
(EDF 2008). Emissions from Natural Gas Production in the Barnett Shale Area and
Opportunities for Cost Effective Improvements. Conducted by Department of
Environmental and Civil Engineering, Southern Methodist University, for Environmental
Defense Fund. Peer-Review Draft. September 30, 2008.

(KSU 2011). Final Report: Cost-Effective Reciprocating Engine Emissions Controls and

Monitoring for E&P Field and Gathering Engines. K. Hohn and S. Nuss-Warren, Kansas
State University. November 2011.

(OTC 2012). Technical Information Oil and Gas Sector, Significant Stationary Sources ofNOx
Emissions. Final. October 17, 2012. [This document focuses on Offshore Gulf of
Mexico, Rocky Mountains, Southwest, and Mid-Continent areas.]

(ETCG 2013). Gas Compressor Engine Study for Northeast Texas, for East Texas (\>uncil of

Governments. Prepared by ENVIRON International Corporation, for East Texas Council
of Governments. June 2013.

(CO DPHE). Reciprocating Internal Combustion Engine (RICE) Source Category, Reasonable
Progress Evaluation for RICE Source Category. Colorado Department of Public Health
and Environment—Air Pollution Control Division.

E-ll


-------
EPA Question 3: What is the effect of NOx SIP call controls on RICE in NOx SIP call
states? That is, what percent reduction and types of controls have gone into place in states
affected by the NOx SIP call?

Notes for Question 3

The applicability, reduction achieved, and cost for RICE NOx controls are often engine
specific and highly variable. (DE 2012) (OTC 2012)

Common NOx control techniques are provided in the table below, along with \()k
emission reductions achievable. (References from other areas outside of the NOx SIP call states
also provided details on controls and emissions reductions achieved by these controls and are
included in the table.)

Effectiveness of Combustion Control Technologies and Add-On Controls

Control Technique

(OTC 2012)

(KSU 2011)

(CARB 2001) (CO DPHE)

2 Stroke, LB

Improved combustion air
flow, Turbocharger

Retard ignition timing

Improved air fuel mixing.
High Pressure Fuel Injection

Screw-in PCC
Autobalance cylinders

(p. 18, 31): up to
75%

(p. 54): diescl. 10%
(reduces engine
efficiency: increases
PM)

(p. 18. 31): up to

90%

Up to 90%:
0.5 to 2 g/bhp-hr
(increases fuel
economy: may
increase CO) (p. 9)

Up to 10% (increase
fuel economy: may
increase CO) (p. 9)

Advanced In-cvlindcr mixing —

Precombuslion chamber
(PCC) ignition system

Micro Precombuslion
chamber (MPCC). hybrid of
Highenergv Ignition system
and PCC

(p. 19, 31-32): up to
90%

(p. 23): not
provided

(p. B-7.8): 15 to	20% (pp. 5-7);

30% (increases	$310 to

fuel consumption;	$2,000/ton

increases VOC,	(p. 8)

HAP)

30 to 70% (p. 11) —

1	g/bhp-hr (p. 10) —

2	to 4 g/bhp-hr (p. 10) —

1 g/bhp-hr (p. 10) —

(continued)

E-12


-------
Effectiveness of Combustion Control Technologies and Add-On Controls (continued)

Control Technique

(OTC 2012)

(KSU 2011)

(CARB 2001) (CO DPHE)

2 Stroke, LB (cont.)

Air to Fuel Ratio Controller
(AFRC)

(p. 19, 32): not
provided

Combustion modifications,
Layered Combustion controls

(p. 25): 60 to 90%;
range of 0.5 to 3
g/bhp-hr

Not provided; use in
combo with Increased
air flow, or
postcombustion
Catalyst; a few
thousand $ for small
engine to $30K for
larger engines (p. 12).

(p. B-8): not
provided (fuel
consumption
penalty of 3%;
may increase CO.
VOC)

5 to 30%
(pp. 5-7); $320
to $8,300/ton
(P- 7)

4 stroke, LB

EGR and NSCR

Combustion modifications,
Layered Combustion controls

(p. 32): (emissions
lower than SCR)a

(p. 38): 90%; range
of 0.5 to 2 g/bhp-hr

Engines (general) or LB

High energy ignition system
(HEIS)

Low emission combustion
(LEC)/precombustion
chamber retrofit (PCC) [also
applicable to RB]

Turbochaiuinu'
superchaiuiiiu. and
Aftercoolniu

(p. 18, 31,44) lu",

2 5 In ' u hhp-hr
ipp 'J-1 ID

ip IS) I p in "5".. —

EGR

ip 55) diesel.
-40% (lobb of fuel
efficiency; loss of
engine output)

Still under
development for NG
engines; not cost
effective at this time
(p. 11).

Ignition system improvement —

(p. B-12): 200
ppm NOx

(p. B-10): 80%
(may increase
VOC, CO)

(p. B-13): 3 to
35% for
Aftercooling
(may reduce
VOC, CO;
increases engine
efficiency, power
rating)

(p. B-14): 30%
(reduces engine
peak power;
reduces fuel
efficiency by 2 to
12%)

(p. B-ll-2): not
provided (may
increase VOC,
CO)

(continued)

E-13


-------
Effectiveness of Combustion Control Technologies and Add-On Controls (continued)

Control Technique

(OTC 2012)

(KSU 2011)

(CARB 2001) (CO DPHE)

Engines (general) or LB
(cont.)

Homogeneous charge
compression ignition
(HCCI), combines best
features of SI and CI engines

Fuel switching,
Hydrogen/NG blended fuel

Selective Catalytic Reduction
(SCR)

Lean-NOx catalysts

(p. 19, 32, 55): 50 to
95% (reduces THC,
CO)

(p. 55): diesel,
50%

10 In

NOxTech

LeanNCK imps

NOx Adsmlvi
(SCOMK)

^"..i(P- 14) (p. B-23): >80% 80 to 90% (pp. 5-7); $430 to $4,900/ton (P- 9) (p. B-24): diesel, — 25 to 50% (increases fuel consumption; may increase VOC, PM) (p. B-25): 80 to — 90%; (decreases CO, VOC, PM by 80%; fuel penalty 5 to 10%) (p. B-27): >90% — on diesel engine <100 hp; [2 ppmv on NG turbine] 50 to 95% (pp. 5-7) (continued) E-14


-------
Effectiveness of Combustion Control Technologies and Add-On Controls (continued)

Control Technique

(OTC 2012)

(KSU 2011)

(CARB 2001) (CO DPHE)

Engines (general) or LB
(cont.)

Fuel switching, methanol —

Hybrid system, modification
of dual bed NSCR system

Use of electric motors in
place of combustion engines

(p. B-16): 30%
for conversion
from NG to
methanol (c;m
generate
formaldehyde
emissions)

(p. B-22): 3 in 4
ppm NOx

(p. B-27): >60%

60 to 100%
(pp. 5-7); $100
to $4,700/ton
[not include
Ml costs]

(P- 9)

RB

Nonselective catalytic
reduction (NSCR) plus
AFRC

Convert RB to LB

EGR

Pre-stratified charge
(converts RB to LB)

(p. 45, 49-51): 90 to
99% (reduces CO.
VOC)

>90%. < 1 g/bhp-hr
(reduces CO. HC)
(p. 13)

(p. B-19-20):
>90% (reduces
CO >80%:
reduces
CO>5()%:
increases fuel
consumption)

80 to 90%
(pp. 5-7);
Capital cost is
$35,000; O&M
is $6,000;
Annualized
capital is
$4,851; TAC is
$10,851;
$571/ton (p. 8)

80% (improved fuel efficiency) E-15


-------
aSome industry literature suggests that some particular 4S RB SI reciprocating engines can be converted to LB
configurations with the accompanying LB engine NOx reduction capabilities. One vendor indicates that
conversion to a LB configuration and the use of exhaust gas recirculation (EGR) delivers the advantages of a LB
engine's efficiency and the RB engine's capability of utilizing NSCR for NOx control. The ability to convert a RB
engine to a LB configuration is highly unit specific and does appear to have had widespread application in
industry (p. 45). (OTC 2012)

Illinois: IEPA projected 2007 NOx emissions from 28 engines subject to the NOx SIP
call to be 6,618 ton/season. NOx emission reductions from these sources were estimated to be
5,422 ton/season, and controlled NOx emissions levels were estimated to be 1,196 ton/season.
(So baseline emissions were estimated to be 6,618 ton/season and controlled emissions were
estimated to be 1,196 ton/season.) (IEPA 2007)

IEPA 2002 base year emissions inventory was 23,347 tpy NOx emitted from RICE and
turbines, or approximately 8.4 percent of total point source NOx emissions (277,899 tpy NOx
emissions from all point sources in Illinois) (p. 12). (IEPA 2007)

In addition to the NOx SIP Call requirements, IEPA also included additional units in its
NOx regulation. NOx SIP Call units were to comply by May 2007, and additional units in NAA
and AA were to comply in 2009, 2011, and 2012 (p. 5 1). The 1L regulation will potentially affect
202 RICE engines and 36 turbines and reduce NOx emissions by 5,422 ton/season in 2007 ozone
control season (p. 10). (IEPA 2007) [Full implementation of the IL regulation in 2012, to include
additional units in NAA and AA counties down to the 500 hp size [28 NOx SIP Call units plus
an additional 246 engines], was projected to reduce NOx emissions statewide by 17,082 tpy and
7,206 ton/season, which is 65 percent reduction on an annual basis and 55 percent reduction in
O3 season emissions (pp. 1 1 and 56). Uncontrolled NOx emissions in 2012 were projected to be
21,532 tpy and 9,134 ton/season, for those units included under the full implementation of the
rule (p. 56). (TEPA 2007)]

Other A vuilable Information

Additional RB control technologies and data are available in the OTC 2012.

Additional Diesel control technologies and data are available in OTC 2012.

References

(CARB 2001). Determination of Reasonably Available Control Technology and Best Available
Retrofit Control Technology for Stationary Spark-Ignited Internal Combustion Engines.
California Environmental Protection Agency, Air Resources Board, Stationary Source
Division, Emissions Assessment Branch, Process Evaluation Section. November 2001.

E-16


-------
(IEPA 2007). Technical Support Document for Controlling NOx Emissions from Stationary
Reciprocating Internal Combustion Engines and Turbines. AQPSTR 07-01. Illinois
Environmental Protection Agency, Air Quality Planning Section, Division of Air
Pollution Control, Bureau of Air. March 19, 2007.

(KSU 2011). Final Report: Cost-Effective Reciprocating Engine Emissions Controls and

Monitoring for E&P Field and Gathering Engines. K. Hohn and S. Nuss-Warren, Kansas
State University. November 2011.

(DE 2012) Background Information, Oil and Gas Sector, Significant Sources ofAY h Emissions.
Delaware Department of Natural Resources and Environmental Quality.

(OTC 2012). Technical Information Oil and Gas Sector, Significant Stationary Sources of NOx
Emissions. Final. October 17, 2012.

(CO DPHE). Reciprocating Internal Combustion Engine (RICE) Source Category, Reasonable
Progress Evaluation for RICE Source Category. Colorado Department of Public Health
and Environment - Air Pollution Control Division.

E-17


-------
EPA Question 4: What are typical or realistic baseline and controlled NOx emissions
factors (grams/hp-hr) for RICE in the OTC states?

Notes for Question 4

NOx control requirements for several of the Ozone Transport Commission (OTC) states
were provided for Connecticut, New Jersey, New York, and Rhode Island, based on a 1994
STAPPA/ALAPCO document (p. 45). (IEPA 2007) These could potentially be used as
maximum EF for RICE units. NOx control requirements are listed in the following table

NOx Control Requirements for RICE in Some OTC States and Other States

State

Covered

NOx Control Level

Reference

Connecticut

>3 MMBtu/hr (1175 hp)

Liquid-fired. CI: 8 g/bhp-hr(584 ppm)

IEPA 2007

New York

RACT for Major
Facilities of NOx, Severe
O3 NAA >200 hp and
Rest of state >400 hp

¦	Thru March 31. 2005. NG. RICE. LB: 3
g/bhp-hr (220 ppm)

¦	Aflcr April 1. 2005. LB: 1.5 g/bhp-hr (110
ppm)

¦	Thru March 31. 2005. Liquid-fired. CI: 9
g/bhp-hr (657 ppm)

¦	Aflcr April 1. 2005: 2.3 g/bhp-hr (168 ppm)

OTC 2012, DE
2012, IEPA 2007

New York
(RACT)

Major facilities >25 tpy,
NYC and Lower Orange
Co: >200 kW

Rest of stele, major
facilities >100 Ipv: >400
kW

¦	NG: 1.5 g/bhp-hr

¦	Landfill or digester gas: 2.0 g/bhp-hr



New Jersey



¦	NG. LB. >500 hp: 2.5 g/bhp-hr (182 ppm)

¦	Liquid-fired. CI, >500 hp: 8 g/bhp-hr (584
ppm)

IEPA 2007

New Jersey
(RACT)

>148 k\\

Group of 2 or more
engines, each al >37 to
<148 kW. but lolal
combined power >148
kW

¦	Gas. LB: 1.5 g/bhp-hr, or 80% reduction

¦	Gas. RB: 1.5 g/bhp-hr



New Jersey
(RACT)

>37 kW

¦	Commenced on or after March 7, 2007:
0.9 g/bhp-hr

¦	Modified on or after March 7, 2007: 0.9
g/bhp-hr, or 90% reduction



Maryland

NG pipeline engines with
>15% capacity factor

NA

IEPA 2007

Other States and Areas





Illinois

NA

¦ 3 g/bhp-hr (210 ppm)

NA

(continued)

E-18


-------
NOx Control Requirements for RICE in Some OTC States and Other States (continued)

State

Covered

NOx Control Level

Reference

Other States and Areas (cont.)

SJVAPCD Rule 4702

(amended ICE SI and CI

2011 Aug 18) nameplate rating >25 hp

Texas

Texas

Texas

Oil & Gas Handling and
Production Facilities

Oil & Gas Handling and
Production Facilities

Texas

(NAA major
sources)

Texas

(NAA minor
sources)

Texas

Oil & Gas Handling and
Production Facilities

RACT. Major ICI. 03
NAA. Beaumont-Port
Arthur O? NAA Major
sources

Combustion Control at
Minor Sources in O3
NAA, Houston-
Galveston-Brazoria

03 NAA, Dallas Ft.
Worth

2S, LB, NG, <100 hp: 75 ppmvd

LB limited use or Gas compression: 65
ppmvd

LB, all others: 11 ppmvd

2S, SI, LB, >500 hp:

-	Mfg before 9/23/1982: 8 g/bhp-hr

-	Mfg before 6/18/1992, <825 hp: 8 g/bhp-
hr

-	Mfg btwn 9/23/1982 and 6/18/1992,
>825hp: 5 g/bhp-hr

-	Mfg btwn 6/18/1992 and 6/1/2010: 2
g/bhp-hr (except 5 g/bhp-hr at reduced
speed and torque 80-100%)

-	Mfg after 6/1/2010: 1 g/bhp-hr

4S, SI. LB:

-	Mfg before 9/23/1982. >500hp: 5 g/bhp-hr
(except 8 g/bhp-hr at reduced speed and
torque 80-100%)

-	Mfg before 6/18/1992. <825 hp: 5 g/bhp-
hr (except 8 g/bhp-hr at reduced speed and
torque 80-100%)

-	Mfg btwn 9/23/1982 and 6/18/1992.

X25hp: 5 g/bhp-lir
\ 1 fg btwn 6/18/1992 and 6/1/2010,
500hp: 2 g/bhp-lir (except 5 g/bhp-hr at
reduced speed and torque 80-100%)

-	Mfg after 6/1/2010, >500hp: 1 g/bhp-hr

After 1/1/2030, no 4S LB SI engine NOx
emissions shall exceed 2 g/bhp-hr regardless
of manufacture date.

4S SI. LB. <500hp:

Mfg before 7/1/2008: 2 g/bhp-hr
After 1/1/2030: no 4S LB SI engine NOx
emissions shall exceed 2 g/bhp-hr regardless
of manufacture date.

NG, SI, RICE, LB >300 hp: 3 g/bhp-hr
NG, SI, RICE, RB, >300 hp: 2 g/bhp-hr

NG, RICE, >50 hp: 0.5 g/bhp-hr

RB, >50 hp: 0.5 g/hp-hr
LB, >50 hp:

-	Installed or moved before June 2007: 0.7
g/hp-hr

-	Installed or moved after June 2007: 0.5
g/hp-hr

OTC 2012

O TC 2012

OTC 2012

OTC 2012

OTC 2012, DE
2012

DE 2012, ETCG
2013

EDF 2008;
ETCG 2013

(continued)

E-19


-------
NOx Control Requirements for RICE in Some OTC States and Other States (continued)

State

Covered

NOx Control Level

Reference

Other States and Areas (cont.)

Texas	East Texas Combustion

Rule (existing engines
comply by March 1,
2010; new engines
comply at startup.)

Colorado

USEPAPart
60, subpart
JJJJ (NSPS)
(final

2008Janl8)

USEPAPart
60, subpart
JJJJ (NSPS)
(final

2008Janl8)

USEPAPart
60, subpart
JJJJ (NSPS)
(final

2008Janl8)

USEPA Pari
60, subpart
JJJJ (NSPS)
(final

2008Janl8)

Regulation 7, RICE, LB,
NG, New, modified,
relocated

NG, SI, ICE

SI, NG and SI, LB, LPG,
100 to 500 hp

NGandLK.. SI. I.li.
500 to 135u hp

SI. Mi and SI. I.li. I.l>(,
(cxccpi I.li 5iio to 1350
hpi

RB, NG, RICE, 240 to 500 hp: 1 g/hp-hr ETCG 2013

RB, NG, RICE, >500 hp: 0.5 g/hp-hr

RB, Landfill gas, RICE, >500 hp: 0.6 g/hp-hr

After July 1, 2007, >500 hp: 2 g/bhp-hr	()|(' 2u 12: CO

After July 1, 2010, >500 hp: 1 g/bhp-hr	I>1*1 If

After January 1, 2008, 100 to 500 hp: 2 g/bhp-
hr

After January 1, 2011, 100 to 500 hp: 1 g/bhp-
hr

Mfg after 7/1/2008. <25 hp. Class I: 11.0 ETGC2013
g/hp-hr of NMHC + NOx combined

Mfg after 7/1/2008. <25 hp. Class I-B: 27.6
g/hp-hr of NMHC + NOx combined

Mfg after 7/1/2008. <25 hp. Class II: 8.4 g/hp-
hr of NMHC + NOx combined

Mfg after 7/1/2008. 25 to 100 hp: 2.8 g/hp-lir
of HC + NOx combined

Mfg after 7/1/2008: 2 g/bhp-hr	ETCG 2013

Mfg after 1/1/2011: 1 g/bhp-hr

\ 1 lu after 7/1/2008: 2 g/bhp-hr	OTC 2012

MI'u after 7/1/2010: 1 g/bhp-hr

\llu after 7/1/2007: 2 g/bhp-hr	ETCG 2013

E-20


-------
NOx Control Requirements for RICE in Local Areas.

State or





Area

Criteria

NOx control level Reference

SCAQMD

Rule 1110.2 Emissions

¦ >500 hp: 0.5 g/bhp-hr (36 ppmvd) OTC 2012

(July 2010)

from Gaseous and Liquid

¦ <500 hp: 0.6 g/bhp-hr (45 ppmvd)



Fueled Engines

¦ After July 1, 2010, >500 hp: 0.15 g/bhp-hr





(11 ppmvd)





¦ After July 1, 2010, <500 hp: 0.6 g/bhp-hr





(45 ppmvd)





¦ After July 1, 2011, All: 0.15 g/bhp-hr (1 1





ppmvd)

For engines with unknown pre-rule emissions, NOx emissions were assumed to be
16.4 g/bhp-hr for 2S and 18.9 g/bhp-hr for 4S. (DE 2012)

A list of Stack test results for engines in PA that are >500 hp are given in Appendix A,
Table 3 of the PA DEP 2013 reference (p. 53). (PA DEP 2013) [Capital] costs for NSCR, RB
ranged from $10 to $12/bhp. (p. 9) NSCR, RB ranged from $ 10 to $ 15/bhp (slightly different
value given here), (p. 16) (MECA 1997)

IC Engine Typical Emissions Levels (MECA 1997)

Engine Type

Lambda (Actual A/F ratio to
Stoichiometric A/F ratio)

Mode

NOx, g/bhp-hr

NG

0.98

Rich

8.3



0.99

Rich

11.0



1.06

Lean

18.0



1.74

Lean

0.7

Diesel

1.6-3.2

Lean

11.6

Dual I'ucl

1.6-1.9

Lean

4.1

For RB, CARB 2001 document has Costs for NSCR w/o AFRC achieving 96%
reduction. Capital costs ranged from $11,000 to $44,000; Annual costs ranged from $8,200 to
$18,000; and cost effectiveness ranged from $2,100/ton to $300/ton NOx reduction (p. V-2 to
V-3). (CARB 2001)

For RB, CARB 2001 document has Costs for Pre-stratified Charge, achieving 80%
reduction. Capital costs ranged from $10,000 to $47,000; Annual costs ranged from $2,700 to

E-21


-------
$11,000; and cost effectiveness ranged from $800/ton to $200/ton NOx reduction (pp. V-2 to V-
3). (CARB 2001)

CARB 2001 document has costs for Ignition Timing Retard (ITR), although the
description of the combustion technology indicates it is less popular on Stationary engines than
mobile source engines (pp. V-2, B-7 to B-8). (CARB 2001)

The EDF 2008 reference provided NOx EF for engines in the Bartlett Shale region. The
document notes that extending the 2009 engine rules in Barnett Shale to counties outside the
DFW NAA would likely result in many engine operators installing NSCR on RB engines. NSCR
costs were cited as follows: $330/ton (IEPA 2007); $92 to $105/ton (EPA 2006); and $ 112 to
$183/ton (northeast Texas 2005 report). Another control technique reviewed in this report
included replacement of compressor engines with electric motors. There are multiple
compressors driven by electric motors throughout Texas (p. 26). Use of electric motors instead of
gas-fired engines eliminates combustion emissions (p. 27). The costs are time and site specific,
based on the cost of electricity, cost of NG, hours of operation per year, number of compressors,
size of compressor, etc. (EDF 2008)

NOx Emission Factors for Engines Identified in DFW 2007 Engine Survey (EDF 2008)

2007 EF

2009 EF

Engine
Type

Engine
Size, hp

NOx, g/hp-hr

Engine Type

Engine Size, hp

NOx, g/hp-hr

RB

<50

13.6

RB

<50

13.6

RB

5o son

13.6

RB

50-500

0.5

RB

5()()

0.9

RB

>500

0.5

LB

5()()

6.2

LB, installed or
moved before June
2007

<500

0.62

LB

5()()

0.9

LB, installed or
moved after June
2007,

<500

0.5

	

	

	

LB, installed or
moved before June
2007

>500

0.7

	

	

	

LB, installed or
moved after June 2007

>500

0.5

E-22


-------
References

(MECA 1997). Emission Control Technology for Stationary Internal Combustion Engines:
Status Report. Manufacturers of Emission Controls Association (MECA). July 1997.

(CARB 2001). Determination of Reasonably Available Control Technology and Best Available
Retrofit Control Technology for Stationary Spark-Ignited Internal Combustion Engines.
California Environmental Protection Agency, Air Resources Board, Stationary Source
Division, Emissions Assessment Branch, Process Evaluation Section. November 2001.

(IEPA 2007). Technical Support Document for Controlling NOx Emissions from Stationary
Reciprocating Internal Combustion Engines and Turbines. AQPSTR 07-01. Illinois
Environmental Protection Agency, Air Quality Planning Section, Division of Air
Pollution Control, Bureau of Air. March 19, 2007.

(EDF 2008). Emissions from Natural Gas Production in the Bar net! Shale Area and
Opportunities for Cost Effective Improvements. Conducted by Department of
Environmental and Civil Engineering, Southern Methodist University, for Environmental
Defense Fund. Peer-Review Draft. September 30, 2008.

(DE 2012) Background Information, Oil and Gas Sector, Significant Sources of NOx Emissions.
Delaware Department of Natural Resources and Environmental Quality.

(OTC 2012). Technical Information Oil and (!as Sector, Significant Stationary Sources of NOx
Emissions. Final. October 17, 2012.

PA DEP 2013. Technical Support Document General Permit GP-5. Pennsylvania Department of
Environmental Protection, Bureau of Air Quality. January 31, 2013.

(ETCG 2013). Gas Compressor Engine Study for Northeast Texas, for East Texas Council of

Governments. Prepared by ENVIRON International Corporation, for East Texas Council
of Governments. June 2013.

(CO DPHE). Reciprocating Interna! (\>m bust ion Engine (RICE) Source Category, Reasonable
Progress Evaluation for RI( E Source (\itegory. Colorado Department of Public Health
and Environment—Air Pollution Control Division.

E-23


-------
EPA Question 5: Using FERC data or other data sources, what is the relationship between
RICE model and age, and emissions (both for baseline and with controls)? In particular,
what is the relationship for RICE built before the imposition of the SI (spark ignition,
natural gas-fired) RICE NSPS in 2007?

Notes for Question 5

The DE 2012 reference stated that many of the installed mainline NG compressors are of
the age (in excess of 40 years old) to have pre-dated modern original equipment manufacturer
(OEM) installed NOx emission controls and otherwise applicable new source performance
standards (NSPS). There is little information on the number of units that may have undergone
NOx modifications as a result of federal or State rules and regulations. The reference cited a
2003 Pipeline Research Council International (PRCI) document that identified 5,686 engines:
71% are LB and 29% are RB (based on dropping the turbine numbers in the table below). The
average age for each unit type is shown in the following table. [These data are repeated in OTC
2012.] [Based on these data, it is estimated that the LB and RB engines are 37 years old on
average (based on dropping the turbine numbers in the table below).] (p. 19) (DE 2012)

2003 Pipeline Research Council International Data (PRC!)

Unit Type	U.S Total Units (%)	Average Ajjc (as of 2003)	Avg hp

2SLB	2.955 (44%)	42	2,113

4SLB	1.059 (16%)	33	1,844

RB	1.672 (25%)	32	589

Turbine	1.016(15%)	24	6,121

The OTC 2<) 12 reference indicated that many of the reciprocating engines driving
mainline NG compressors are in excess of 40 years old, pre-dating any applicable modern OEM
installed NOx emission control and any otherwise applicable NSPS NOx controls (p. 16). (OTC
2012)

The DE 2012 reference discussed a 2005 study conducted for NG field gathering engines
in Eastern Texas; the study was able to determine the age only for a very small portion of the
engines, and the engine age ranged from 2 to 25 years. The output ratings of engines in the study
ranged from 26 to 1478 hp, with the majority rated between 50 and 200 hp (p. 12). (DE 2012)

The DE 2012 reference indicated they reviewed MARAMA's 2007 Point Source
Inventory and 2007 FERC data. The 2007 FERC data are provided as Attachment III to the

E-24


-------
reference. The two sets of data did not match: 2007 MARAMA data indicated 107 compressor
facilities, and 2007 FERC data indicated 150 compressor facilities. The reviewed databases did
not provide any information regarding NOx emission rates (g/bhp-hr, ppmvd). NOx emission
rates were obtained for a small number of prime movers, through operating permits: 2SLB range
from 1 to 13.3 g/bhp-hr; 4SLB range from 0.5 to 6 g/bhp-hr; and 4SRB were 3 g/bhp-hr. The
data are not sufficient to estimate actual NOx emission rates and NOx reductions. Note that the
FERC data addresses large entities, and smaller companies may not be required to report data to
FERC. The 2007 OTC compressors from FERC are provided in the following table. (DE 2012)

State

No. Compressors

Total Rated hp

CT

10

35,300

MA

15

25,702

MD

17

52.250

ME

4

33.244

NJ

36

129.130

NY

120

359.487

PA

4(.~

1.131,164

RI

o

29,170

VA (OTR area only)

22

49,390

The KSU 2011 reference discussed control technologies testing performed in the
laboratory on a 1966 Ajax DP-1 15 (Lean Burn) that has none of the low emissions controls that
are currently OEM standard. The published emission factor (EF) for this engine is 4.4 g/bhp-hr,
and the emissions from actual testing were 4.69±0.18 g/bhp-hr (the Lab testing results are
discussed on pp. 19-27). There is additional discussion of Field testing conducted on multiple LB
engines with NOx emission control techniques, including (1) Increased air flow, and
precombustion chamber (PCC) screw-in type, (2) PCC screw-in type and Upgraded
turbocharger, (3) Integral PCC and high-output turbocharger (pp. 27-29). Discussion of Field
testing conducted on two RB engines with NOx emission control techniques (p. 29). Integrated
nonselective catalytic reduction (NSCR) with modeling and enhanced controller is also
discussed. (KSU 2011).

References

(KSU 2011). Final Report: Cost-Effective Reciprocating Engine Emissions Controls and

Monitoring for E&P Field and Gathering Engines. K. Hohn and S. Nuss-Warren, Kansas
State University. November 2011.

E-25


-------
(DE 2012) Background Information, Oil and Gas Sector, Significant Sources ofNOx Emissions.
Delaware Department of Natural Resources and Environmental Quality.

(OTC 2012). Technical Information Oil and Gas Sector, Significant Stationary Sources ofNOx
Emissions. Final. October 17, 2012.

E-26


-------
EPA Question 6: What is the variability in NOx emissions from RICE within each State,
both for baseline and with controls?

Notes for Question 6

No data were found, r Likely a review of RICE SCCs in the NEI across states would be a
useful exercise to see the relative levels of baseline and/or controlled NOx emissions, however
this exercise was not part of this task.1

E-27


-------
From:

Subject:

Date:

To:

US EPA OAQPS
SRA International, Inc.

Review of CoST Model Emission Reduction Estimates

September 30, 2014

EPA uses the Control Strategy Tool (CoST) to estimate the emission reductions and engineering costs
associated with control strategies applied to point, area, and mobile sources of air pollutant emissions to
support the analyses of air pollution policies and regulations. CoST accomplishes ill is by matching
control measures to emission sources using algorithms such as "maximum emissions reduction", "least
cost", and "apply measures in series". There was a concern that the baseline m\ eniorv used by CoST did
not completely account for emission control requirements already in place, and lhal llic emission
reductions were perhaps overestimated.

SRA reviewed the CoST results and made recommeiidalions lor chaiiL:iny llie ( oST control measure
assignment and the estimated reductions for o\ides ol'nilrogen (\(K) The recommendations were based
on a review of source permits, state regulations, enforcement actions. and oilier available information.
The analysis was conducted for a 24-state area in the eastern two-lhirds of the U.S. The focus was on
stationary point sources other than electric generating units (non-EGUs). The purpose of this memo is to
document the data used and assumptions made in recommending changes to the CoST results, and to
summarize the differences Ivlxxeen llie CoST results and the recommended changes.

The findings in this memo are Ixised on rex ie\x of CoST results for a 2018 emissions inventory projected
from the 2011 National Lmission Inxenlorx (\LI). This work was in support of EPA's current Transport
Rule efforls lor implementing llie 75 pph ozone standard. If EPA considers establishing a tighter ozone
standard in llie lulu re. il is likelx dial a more dislant future year will be used and that some of the
conclusions reached in lliis memo could change.

CoST DATA PROVIDED BY EPA

EPA provided SRA xvith the outputs from a CoST scenario that identified sources for which NOx controls
were available at a cost-effectiveness level of less than $10,000 per ton. The CoST outputs included
source identifiers, control technology, baseline emissions and estimates of NOx emission reductions. The
CoST results were divided into two groups. The first group included sources where CoST estimated NOx
emission reductions of more than 100 tons per year. There were 547 sources in this group where CoST
controls were initially applied. The second group included sources where CoST estimated emission
reductions for sources whose 2018 projected emissions were greater than 25 tons/year, excluding those

1


-------
with reductions greater than 100 tons/year. There were 1,280 sources in this group where CoST controls
were initially applied.

Another contractor reviewed the CoST results for additional source categories, and their
recommendations were merged with SRA's recommendations in the summary tables and maps that
follow. The data used, assumptions made and results for IC engines are documented elsewhere1.

REVIEW OF CoST RESULTS FOR THE GREATER THAN 100 TI>Y GKOl l»

Table 1 summarizes the source categories included in our analysis, the CoST recommendation for NOx
control, and the recommendation for changing the CoST control measure assignment and associated
emission reduction estimates. Following Table 1. there is a discussion for each source group to provide
more detail on the rationale for the recommended changes for each source group. Attachments 1 to 4 are
tabular comparisons of the initial CoST emission reduction estimates and the recommended changes. All
Attachments present the results in terms of tons per ozone season, simply estimated by assuming that
ozone season emissions were equal to 5/12 of the annual emissions. Maps 1A and 1B graphically show
the location of sources and the magnitude of the recommended emission reductions.

Table 1 - CoST Controls and Recommended Changes for
Greater than 100 TPY Sources

Source Group

CoST Control
Recommendation

Summary of Recommended Changes to
CoST Controls and Reductions

Ammonia - NG-fired
Reformers

Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a permit or
state regulation already required the source
to be controlled.

By-Product Coke Mfg;
Oven Underfiring

Selective Non-Catalytic
Reduction

Review of a source-specific NOx RACT
permit indicated that NOx controls were
technically or economically infeasible.

Cement Kilns

Biosolid Injection
Technology

Disagreed with CoST recommendation
based on concerns about biosolids
availability and information from EPA's ISIS
(Industrial Sector Integrated Solutions)
Model; recommended SNCR for all sources,
except those that already have SNCR due to
NOx SIP Call, NSR requirement, Consent
Decree, or other state regulation.

1 Update of NOx Control Measure Data in the CoST Control Measures Database for Four Industrial
Source Categories: Ammonia Reformers, NonEGU Combustion Turbines,Glass Manufacturing, and Lean
Burn Reciprocating Internal Combustion Engines," prepared by Research Triangle Institute, July 2014.

2


-------
Source Group

CoST Control
Recommendation

Summary of Recommended Changes to
CoST Controls and Reductions

Cement Manufacturing
- Dry

Selective Non-Catalytic
Reduction

Agreed with CoST recommendation except
when already controlled due to NOx SIP Call,
NSR requirement, Consent Decree, or other
state regulation.

Cement Manufacturing
-Wet

Mid-kiln Firing

Disagreed with CoST recommendation
based on information from EPA's ISIS Model;
recommended SNCR for all sources, except
those that already controlled

Coal Cleaning -
Thermal Dryer

Low NOx Burner

Agreed with CoST recommendation

Comm/lnst Incinerators

Selective Non-Catalytic
Reduction

Both sources are already controlled with
SNCR

External Combustion
Boilers, ElecGen, Solid
Waste

Selective Non-Catalytic
Reduction

All 6 sources are already controlled with
SNCR

Fluid Catalytic Cracking
Units

Low NOx Burner and Flue
Gas Recirculation

Nearly all FCCUs are already controlled due
to the OECA global refinery consent decrees.
There is one small refinery in West Texas
that does not appear to be covered by a
consent decree, so the CoST
recommendation was accepted.

Glass Manufacturing -
Container. Flat.
Pressed

OXY-Firing

Disagreed with CoST recommendation.
OXY-firing is not generally required under
recent OECA consent decrees. More
common control is oxygen-enriched air
staging (OEAS). OXY-firing can only be
implemented at the time of furnace rebuild,
which is generally done every 10-15 years.
Changed recommended control to OEAS
with a 50% NOx reduction instead of OXY-
firing at 85% NOx reduction, except for
sources that already had NOx controls in
place due to a consent decree, NSR
requirement, or state regulation. Assumed
that a furnace with a NOx emission limit of
less than 4 lbs/ton of glass pulled was
already reasonably controlled.

ICI Boilers -
Coal/Cyclone

Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a permit or
state regulation already required the source
to be controlled. LADCO/OTC also
recommends SCR

ICI Boilers -
Coal/Stoker

Selective Catalytic
Reduction

Disagreed with CoST recommendation of
SCR. CoST has $2200/ton, which appears

3


-------
Source Group

CoST Control
Recommendation

Summary of Recommended Changes to
CoST Controls and Reductions





very low for ICI boilers. Used LADCO/OTC
recommendation of SNCR for Coal-Stokers
with a 50% reduction, except for those
sources where a permit or state regulation
already required the source to be controlled.

ICI Boilers - Coal/Wall

Low NOx Burner and
Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a permit or
state regulation already required the source
to be controlled. LADCO/OTC also
recommends LNB/SCR

ICI Boilers - Gas,
Natural Gas, Process
Gas

Selective Catalytic
Reduction

Disagreed with CoST recommendation of
SCR. CoST has $3456/ton, which appears
very low for ICI boilers. Used LADCO/OTC
recommendation of Low NOx Burners plus
Flue Gas Recirculation for Gas-fire ICI
boilers with a 60% reduction, except for
those sources where a permit or state
regulation already required the source to be
controlled

Industrial Incinerators

Selective Non-Catalytic
Reduction

Agreed with CoST recommendation of SNCR
except for those sources where a permit or
state regulation already required the source
to be controlled.

Iron & Steel Mills -
Reheating

Low NOx Burner and Flue
Gas Recirculation

Agreed with CoST recommendation except
for those sources where a permit or state
regulation already required the source to be
controlled.

Municipal Waste
Combustors

Selective Non-Catalytic
Reduction

Agreed with CoST recommendation of SNCR
except for those sources where a permit or
state regulation already required the source
to be controlled.

Nitric Acid
Manufacturing

Nonselective Catalytic
Reduction

Agreed with CoST recommendation of NSCR
except for those sources where a permit or
state regulation already required the source
to be controlled.

Petroleum Refinery
Process Heaters

SCR-95%

Nearly all refineries are already controlled
due to the OECA global refinery consent
decrees, which generally require 40-60%
reductions across all boilers/heaters that
each company operates. Not possible at
present to identify the individual
boilers/heaters that actually have been
controlled or are scheduled to be controlled
due to confidentiality agreements between
EPA and companies.

4


-------
Source Group

CoST Control
Recommendation

Summary of Recommended Changes to
CoST Controls and Reductions

Taconite Ore
Processing - Induration
- Coal or Gas

Selective Catalytic
Reduction

Disagree with CoST recommendation of
SCR. EPA Region V considers SCR/SNCR
to be infeasible. Used Low NOx Burners at
70% reduction instead as reasonable control,
except for those sources where a permit or
state regulation already required the source
to be controlled. .

Utility Boilers* -
Coal/Wall

Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a permit or
state regulation already required the source
to be controlled.

Utility Boilers* - Oil/Gas

Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a permit or
state regulation already required the source
to be controlled.

The utility boilers included in the context of this report are non-IPM utility boilers. In the NEI, these units
have an SCC of 1-01—xxx-xx (the SCC series generally used for electric generating units. However, the
sources included in this analysis do not sell electricity to the grid.

Ammonia - NG-fired Reformers

There are 15 sources in this category The CoST conlrol kvhnologv was selective catalytic reduction
(SCR) with a 90% reduction in NOx emissions. We determined that four of these sources were already
controlled by either SCR or ultra-NO\ burners and recommended no further control/reductions. For all
other sources, we agreed with the CoST control and emission reduction estimate.

By-Product Coke Mfg; Oven Underfiring

There are 14 sources in llus calegor\ The CoST control technology was selective non-catalytic reduction
(SNCR) \\ iill a Mi",, ivduclion in \(K emissions. We reviewed a detailed RACT analysis for a facility in
Pennsyh uniu iluil delernuned llml no controls were feasible. For all sources in this category, we
recommended that no controls were feasible and thus no reductions were appropriate.

Cement Preheater/Precalciner Kilns

There are 36 sources in this category. The CoST control technology was biosolid injection technology
with a 23% reduction in NOx emissions. We reviewed permits and consent decrees to identify those kilns
that are already controlled. Several kilns are already controlled based on NOx SIP Call requirements that
typically required low NOx burners, mid-kiln firing, or an approved alternative that resulted in a 30%
reduction. Other kilns already had SNCR installed due to a consent decree, new source review
requirement, or other state-level requirement.

5


-------
EPA expressed a concern whether there was sufficient biosolids availability for use by the uncontrolled
kilns. Also, EPA has done considerable research on cement kiln NOx controls as part of its Industrial
Sector Integrated Solutions (ISIS) project. EPA uses the ISIS-cement model help analyze policy options
for various rulemakings. Based on the ISIS work, we recommended that low-NOx burners and SNCR as
the appropriate control for all types of kilns.

For uncontrolled kilns, we applied a 65% reduction in NOx emissions. For kilns already controlled with
low-NOx burners or mid-kiln firing, we applied a 35% incremental reduction to account for the additional
reductions from SNCR. For kilns already controlled with SNCR, we applied no addilional emission
reductions.

Cement Manufacturing - Dry Process

There are 20 sources in this category. The CoST control technology was SNCR w illi a 5<)"„ reduction in
NOx emissions. We reviewed permits and consent decrees to identity those kilns that are already
controlled. Several kilns are already controlled based on NOx SIP Call requirements that typically
required low NOx burners, mid-kiln firing, or an approved alternative that resulted in a 30% reduction.
Other kilns already had SNCR installed due to a consent decree, new source review requirement, or other
state-level requirement.

As discussed earlier, we recommended that lo\\-M)\ burners and SNCR as the appropriate control for all
types of kilns based on the ISIS work. For unconliolled kilns, we applied a 65% reduction in NOx
emissions. For kilns already controlled with low-\()\ burners or mid-kiln firing, we applied a 35%
incremental reduction to account for the additional reductions from SNCR. For kilns already controlled
with SNCR, we applied no additional emission reductions.

Cement Manufacturing - Wet Process

There arc seven sources in this caleijoiy The CoST control technology was mid-kiln firing with a 30%
reduction in MK emissions \\ e deleinuned lhat two of these kilns were installing a pilot SCR system as
partofaconsenl decree One kiln recently went through NSR review and has state-of-the-art control.
Another kiln is required in install SNCR as part of a consent decree. No additional reductions were
applied for these kilns. For the remaining kilns, we applied low-NOx burners and SNCR as described in
the previous sections
Coal Cleaning - Thermal Dryer

There was one source in this category. The CoST control technology was a low-NOx burner with a 50%
reduction in NOx emissions. We could not find any information on this source and accepted the CoST
controls.

6


-------
Comm/Inst Incinerators

There are two sources in this category. The CoST control technology was SNCR with a 45% reduction in
NOx emissions. Both of these sources are already controlled by SNCR and we applied no additional
emission reductions.

External Combustion Boilers, Elec Gen, Solid Waste

There are six sources in this category. The CoST control technology was SNCR with a 50% reduction in
NOx emissions. All six of these sources are already controlled by SNCR and we applied no additional
emission reductions.

Fluid Catalytic Cracking Units

There are six sources in this category. The CoST control technology was lou-MK burners and flue gas
recirculation with a 55% reduction in NOx emissions. Nearly all sources arc a I read \ con i rolled or
required to install controls as a result of the ERA's global relineiy consent decrees There is one small
refinery in West Texas that does not appear to be covered b\ a consent decree, so llie CoST
recommendation was accepted.

Glass Manufacturing - Container, Flat, Pressed

There are 65 sources in this category. The CoST control technology was o\\ -11 ring with an 85%
reduction in NOx emissions. There were several concerns about using oxv-firing for this analysis. First,
there is a concern about the timing of installing oxv-firing technology. Oxv-firing is typically installed at
the time of a furnace rebuild, which is typically done every 10 to 15 years. Second, oxy-firing is not
generally required under recent EPA consent decrees. More common control is oxygen-enriched air
staging (OEAS). We recommended that OEAS with a 50% NOx reduction instead of OXY-firing at 85%
NOx reduction, except for sources that already had NOx controls in place due to a consent decree, NSR
requirement, or state regulation. We assumed that a furnace w ith a NOx emission limit of less than 4
lbs/ton ol'glass pulled was already reasonably controlled.

ICI Boilers - Coal/C Vclone

There are cighl sources in ill is category. The CoST control technology was SCR with an 80% reduction in
NOx emissions. We reviewed the Evaluation of Control Options for Industrial, Commercial and
Institutional (ICI) lloilers Technical Support Document (TSD), March, 2011 prepared by the Lake
Michigan Air Directors Consortium (LADCO) and the Ozone Transport Commission (OTC).
LADCO/OTC also recommended SCR for coal-cyclone boilers. Since the LADCO/OTC recommendation
was consistent with the CoST control, we agreed with the CoST control technology for five sources
which we determined were uncontrolled. Two sources were determined to be already controlled. One
source appears to have shut down their coal-fired boilers. No reductions were applied for these three
sources since they are already controlled.

7


-------
ICI Boilers - Coal/Stoker

There are 45 sources in this category. The CoST control technology was SCR with an 80% reduction in
NOx emissions. The LADCO/OTC recommendation was for combustion tuning and SNCR. We agreed
with the LADCO/OTC recommendation and assumed a 50% control efficiency. We determined that most
of these sources are currently uncontrolled. Two coal-fired boilers are scheduled to be replaced with gas-
fired boilers. Two other boilers recently installed SNCR.

ICI Boilers - Coal/Wall

There are 54 sources in this category. The CoST control technology was low -\( )\ burners and SCR with
a 91% reduction in NOx emissions. The LADCO/OTC recommendation was also lor low -\()\ burners
and SCR. Since the LADCO/OTC recommendation was consistent with llie ( oST control, we agreed
with the CoST control technology and emission reductions.

ICI Boilers - Gas, Natural Gas, Process Gas

There are 130 sources in this category. The CoST control technology was SCR with an 80% reduction in
NOx emissions. The LADCO/OTC recommendation was for low-NOx burners, flue gas recirculation, or
low-NOx burners combined with flue gas recirculation. We agreed with the LADCO/OTC
recommendation of low-NOx burners combined with flue gas recirculation and assumed a 60% control
efficiency.

Several of these sources arc located in the O'l'R or ozone nonulLiinment areas, and as a result already have
aRACT control requirement or emission limitation that is consistent with the LADCO/OTC
recommendations. A few of ihese sources are located at petroleum refineries and were assumed to be
already controlled due to EPA's ielinei\ enforcement initiative.

Municipal Waste Combustors

There arc 55 sources in this calegoiy The ( oST control technology was SNCR with a 45% reduction in
NOx emissions. We determined that 35 of these sources are already controlled with SNCR and no
additional reductions were applied. For the remaining uncontrolled sources, we agreed with the CoST
controls and emission reductions.

Nitric Acid Manufacturing

There are seven sources in this category. The CoST control technology was non-selective catalytic
reduction (NSCR) with a 98% reduction in NOx emissions. All but one of these sources is already
controlled by NSCR or SCR.

Petroleum Refinery Process Heaters

There are 28 sources in this category. The CoST control technology was SCR with a 95% reduction in
NOx emissions. All of the sources in this category are covered sources under EPA's global refinery
enforcement initiative. The settlements generally require 40-60% reductions across all boilers/heaters that

8


-------
each company operates. Companies have submitted NOx compliance plans to OECA that identify the
specific sources that have been controlled or are planned to be controlled, along with the technology used.
But it is not possible at present to identify the individual boilers/heaters that actually have been controlled
or are scheduled to be controlled due to confidentiality agreements between EPA and companies. No
additional reductions were included for this category.

Taconite Ore Processing - Induration - Coal or Gas

There are 10 sources in this category. The CoST control technology was SCR with a 90% reduction in
NOx emissions. All of the sources in this category are already subject to I icsl Available Retrofit
Technology (BART) requirements under the Regional Haze program. EPA Region Y determined that
BART is low-NOx burners and agreed that SCR controls are infeasible for indurating furnaces. No
additional reductions were included for this category.

Utility Boilers - Coal/Wall, Oil, Gas

There are 11 sources in this category. The CoST control technology was SCR w ilh a 80 to 90% reduction
in NOx emissions depending on fuel type. All of the sources in llus category appear to be uncontrolled
and we agreed with the CoST control and emission reduction estimate

REVIEW OF CoST RESULTS FOR THE 25 TO 100 TPY GROl P

Due to the large number of sources in llns group, we were not able to review individual permits to
determine whether the individual source was already controlled. Instead, our recommendations were
based on of state regulations, enforcement actions, engineering judgment, and other available information.
We generally assumed that sources located in areas w ith stringent NOx rules are already well controlled
and we assumed that no additional reductions were likely from these sources. This assumption was
generally applied in New Jersey. New York and sources located in the Houston nonattainment area.

Given more 11me. w e would 11ke to have also applied this assumption in other areas with stringent existing
regulations, such as Chicago. Milwaukee, and Baton Rouge. In any future analysis, it would be useful to
examine the stringency of rules that apply strictly to nonattainment areas.

Table 2 summarizes the source categories included in our analysis, the CoST recommendation for NOx
control, and the recommendation for changing the CoST control measure assignment and associated
emission reduction estimates. Following Table 2, there is a discussion for each source group to provide
more detail on the rationale for the recommended changes for each source group. Attachments 5 to 8 are
tabular comparisons of the initial CoST emission reduction estimates and the recommended changes. All
Attachments present the results in terms of tons per ozone season, simply estimated by assuming that

9


-------
ozone season emissions were equal to 5/12 of the annual emissions. Maps 3A and 3B graphically show
the location of sources and the magnitude of the recommended emission reductions.

Table 2 - CoST Controls and Recommended Changes for
25 to 100 TPY Sources

Source Group

CoST Control
Recommendation

Summary of Recommended Changes to
CoST Controls and Reductions

Ammonia - NG-fired
Reformers

Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a permit or
state regulation already required the source to
be controlled.

Cement Kilns

Biosolid Injection
Technology

Because of low emissions, assume that the
kiln is already controlled or have very low
usage which would result in a unreasonably
high cost-effectiveness

Cement Manufacturing
-Wet

Mid-kiln Firing

Because of low emissions, assume that the
kiln is already controlled or have very low
usage which would result in a unreasonably
high cost-effectiveness

Ceramic Clay Mfg;
Drying

Low NOx Burner

Questions about technical feasibility for these
category, assume zero reductions

Coal Cleaning -
Thermal Dryer

Low NOx Burner

Agree with CoST recommendation

Comm/lnst Incinerators

Selective Non-Catalytic
Reduction

Agree with CoST recommendation

External Combustion
Boilers, Elec Gen,
Sub/Bit Coal

Selective Non-Catalytic
Reduction

Agree with CoST recommendation, although
questions as to whether the source is already
controlled or very low usage which would
result in a unreasonably high cost-
effectiveness

Fluid Catalytic Cracking
Units

Low NOx Burner and Flue
Gas Recirculation

Nearly all FCCUs are already controlled due
to the OECA global refinery consent decrees.

Gas Turbines

Low NOx Burners

Agreed with CoST recommendation except for
those sources where a state regulation
already required the source to be controlled.

Glass Manufacturing -
Container, Flat, Pressed

OXY-Firing

Because of low emissions, assume that the
furnace is already controlled or have very low
usage which would result in a unreasonably
high cost-effectiveness

ICI Boilers -
Coal/Stoker

Selective Catalytic
Reduction

Disagreed with CoST recommendation of
SCR. CoST has $2200/ton, which appears
very low for ICI boilers. Used LADCO/OTC
recommendation of SNCR for Coal-Stokers
with a 50% reduction, except for those
sources where a state regulation already
required the source to be controlled.

ICI Boilers - Coal/Wall

Low NOx Burner and
Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a state
regulation already required the source to be

10


-------
Source Group

CoST Control
Recommendation

Summary of Recommended Changes to
CoST Controls and Reductions





controlled. LADCO/OTC also recommends
LNB/SCR

ICI Boilers - Distillate
Oil or Process Gas

Selective Catalytic
Reduction

Because of low emissions, assume that the
boiler is already controlled or have very low
usage which would result in a unreasonably
high cost-effectiveness

ICI Boilers - Natural
Gas

Low NOx Burner and
Selective Catalytic
Reduction

Disagreed with CoST recommendation of
SCR. Used LADCO/OTC recommendation of
Low NOx Burners plus Flue Gas Recirculation
for Gas-fire ICI boilers with a 60% reduction,
except for those sources where a permit or
state regulation already required the source to
be controlled

ICI Boilers - Residual
Oil

Low NOx Burner and
Selective Non-Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a state
regulation already required the source to be
controlled.

Industrial Incinerators

Selective Non-Catalytic
Reduction

Agreed with CoST recommendation of SNCR
except for those sources where a state
regulation already required the source to be
controlled.

Iron & Steel Mills -
Reheating

Low NOx Burner and Flue
Gas Recirculation

Agreed with CoST recommendation except for
those sources where a state regulation
already required the source to be controlled.

Municipal Waste
Combustors

Selective Non-Catalytic
Reduction

Agreed with CoST recommendation of SNCR
except for those sources where a state
regulation already required the source to be
controlled.

Nitric Acid
Manufacturing

Nonselective Catalytic
Reduction

Agreed with CoST recommendation of NSCR
except for those sources where a state
regulation already required the source to be
controlled.

Petroleum Refinery
Process Heaters

SCR or Ultra-Low NOx
Burner

Nearly all refineries are already controlled due
to the OECA global refinery consent decrees,
which generally require 40-60% reductions
across all boilers/heaters that each company
operates. Not possible at present to identify
the individual boilers/heaters that actually
have been controlled or are scheduled to be
control due to confidentiality agreements
between EPA and companies.

Utility Boilers -
Coal/Wall

Selective Catalytic
Reduction

Agreed with CoST recommendation of SCR
except for those sources where a state
regulation already required the source to be
controlled

Utility Boilers - Oil/Gas

Selective Catalytic
Reduction

Because of low emissions, assume
unreasonably high cost-effectiveness for
SCR; use LNB/FGR as reasonable control.

11


-------
Ammonia - NG-fired Reformers

There are seven sources in this category. The CoST control technology was SCR with a 90% reduction in
NOx emissions. For all other sources, we agreed with the CoST control and emission reduction estimate.
Cement Kilns

There are six sources in this category. The CoST control technology was either biosolid injection
technology with a 23% reduction in NOx emissions or mid-kiln firing with a 30% reduction. Because of
the low baseline emissions for these kilns, we assumed that the kilns were already controlled or have low
usage which would result in a very high cost-effectiveness. We determined that no reductions be applied
for these sources.

Coal Cleaning - Thermal Dryer

There are 10 sources in this category. The CoST control technology was a lou-\(K burner \\ilha50%
reduction in NOx emissions. We agreed with the CoST control and emission reduction eslimate.
Commercial/Institutional Incinerators

There are four sources in this category. The CoST control technology was S\CR with a 45% reduction in
NOx emissions. We agreed with the CoST control and emission reduction esl imate.

External Combustion Boilers, Electric Generation, Coal

There are 14 sources in this category. The CoST control technology was SJNCR with a 40% reduction in
NOx emissions. It appears that the sources in this category are low usage spreader stokers. Although there
may be a concern about the cost-cffcctivcncss for these sources, we agreed with the CoST control and
emission reduction estimate.

Fluid Catalytic Cracking Units

There are 21 sources in this category. The CoST control technology was low-NOx burners and flue gas
recirculation with a 55% reduction in NOx emissions. All sources in this category are assumed subject to
existing control i'ci|uii'cmenls resulting from the OECA global refinery enforcement initiative.

Additional l\. eight of the sources are located in the Houston nonattainment area and are likely subject to
stringent controls I-'or these reasons, we assumed no further control or emission reductions for the
FCCUs.

Gas Turbines

There are 438 sources in this category. The CoST control technology was for low-NOx burners with a
68% reduction in NOx emissions. We agreed with the CoST control and emission reduction estimate,
except for those sources located in the OTR and Houston ozone nonattainment area, where we assumed
that these sources already had RACT controls.

Glass Manufacturing - Container, Flat, Pressed

12


-------
There are eight sources in this category. The CoST control technology was SCR with a 90% reduction in
NOx emissions. Because of the low baseline emissions for these furnaces, we assumed that the furnaces
were already controlled and determined that no reductions be applied for these sources.

ICI Boilers - Coal/Stoker

There are 133 sources in this category. The CoST control technology was SCR with a 90% reduction in
NOx emissions. The LADCO/OTC recommendation was for combustion tuning and SNCR. We agreed
with the LADCO/OTC recommendation and assumed a 50% control efficiency.

ICI Boilers - Coal/Wall

There are 11 sources in this category. The CoST control technology was SCR w illi a 90% reduction in
NOx emissions. The CoST control technology was low-NOx burners and S('R w ilh a 91 % reduction in
NOx emissions. The LADCO/OTC recommendation was also for low-NOx burners and SCR. Since the
LADCO/OTC recommendation was consistent w illi I lie ( oST control. we agreed wi ill I lie ( oST control
technology and emission reductions.

ICI Boilers - Natural Gas

There are 376 sources in this category. The CoST coin ml technology was low-NOx burners and SCR
with a 91% reduction in NOx emissions. The l..\l)( () OK recommendation was for low-NOx burners,
flue gas recirculation, or low-NOx burners combined w ilh Hue gas recirculation. We agreed with the
LADCO/OTC recommendation of low-NOx burners combined w ith flue gas recirculation and assumed a
50% control efficiency, except for those sources located in the OTR and Houston ozone nonattainment
area, where we assumed thai lliese sources already had RACT controls.

ICI Boilers - Process Gas

There are 57 sources in this category. The CoST control technology was SCR with a 90% reduction in
NOx emissions Most of these sources are located at petroleum refineries and are assumed subject to
existing control requirements resulting from the OECA global refinery enforcement initiative, or are
located in llie I louslon nonallainment area and arc likely subject to stringent controls. For these reasons,
we assumed no I'urther control or emission reductions.

ICI Boilers - Residual Oil

There are 28 sources in this category. The CoST control technology was low-NOx burner and SNCR with
a 69.5% reduction in NOx emissions. We agreed with the CoST control and emission reduction estimate,
except for those sources located in the OTR and Houston ozone nonattainment area, where we assumed
that these sources already had RACT controls.

Industrial Incinerators

There are 21 sources in this category. The CoST control technology was SNCR with a 45% reduction in
NOx emissions. We agreed with the CoST control and emission reduction estimate, except for those

13


-------
sources located in the OTR and Houston ozone nonattainment area, where we assumed that these sources
already had RACT controls.

Iron & Steel Mills - Reheating

There are 32 sources in this category. The CoST control technology was low-NOx burners and flue gas
recirculation with a 77% reduction in NOx emissions. We agreed with the CoST control and emission
reduction estimate.

Municipal Waste Combustors

There are 25 sources in this category. The CoST control technology was SCR w ilh a 90% reduction in
NOx emissions. RTI identified the sources are already controlled and no additional reductions were
applied for these sources. For the remaining sources, we agreed with the CoST controls and emission
reductions.

Nitric Acid Manufacturing

There are 14 sources in this category. The CoST control technology was \S( R with a 98% reduction in
NOx emissions. We agreed with the CoST control and emission reduction estimate.

Petroleum Refinery Process Heaters

There are 30 sources in this category. The CoST control technology was SCR with a 90-98% reduction or
ultra-low NOx burners with a 30-50% reductions in NOx emissions Most of these sources are located at
petroleum refineries and arc assumed subject to existing control requirements resulting from the OECA
global refinery enforcement initiative, or are located in the Houston nonattainment area and are likely
subject to stringent controls. I-'or these reasons, we assumed no further control or emission reductions.
Utility Boilers - Coal/Wall

There arc three sources in this category. The CoST control technology was SCR with a 90% reduction in
NOx emissions. We agreed with the CoST control and emission reduction estimate.

Utility Boilers - ()il/(ias

There arc 27 sources in this category. The CoST control technology was SCR with a 80% reduction in
NOx emissions The l..\l)( () OTC recommendation was for low-NOx burners or flue gas recirculation.
We agreed with the LADCO/OTC recommendation of low-NOx burners combined with flue gas
recirculation and assumed a 60% control efficiency.

14


-------
Attachment 1 - NOx Emission Reductions by State for Sources in the >100 Ton per Year Reduction Group







Recommended







NOx emissions

NOx emissions

Size of correction





reduced from

reduced from

in NOx emission





controls in CoST

controls in CoST

reductions



Number of

(tons/03 season)

(tons/03 season)

(tons/03 season)

State

Sources

(A)

(B)

(A-B)

Alabama

24

2.855

2.287

568

Arkansas

6

455

293

162

Delaware

2

206

0

206

Florida

20

2.158

1.370

788

Illinois

21

2.659

1.472

1,187

Indiana

41

5.405

4.510

896

Iowa

10

1.226

999

227

Kansas

7

735

452

283

Kentucky

11

915

838

77

Louisiana

57

7.623

3.622

4,000

Maryland

10

1.933

355

1,578

Michigan

27

2.758

1,768

990

Mississippi

7

1.054

516

538

Missouri

15

1.698

1,562

136

New Jersey

15

417

0

417

New York

30

3.091

281

2,810

Ohio

37

4.098

2,039

2,058

Oklahoma

20

2.949

1,864

1,086

Pennsylvania

52

5.637

2,215

3,422

Tennessee

13

4,741

1,987

2,755

Texas

65

8,860

6,383

2,477

Virginia

28

3,337

3,033

303

West Virginia

9

1,180

793

387

Wisconsin

20

4,092

3,416

676



547

70,082

42,054

28,028

15


-------
Attachment 2 - NOx Emission Reductions by Source Group for Sources in the >100 Ton per Year Reduction Group

Source Group

Ammonia - NG-Fired Reformers
By-Product Coke Mfg; Oven Underfiring
Cement Kilns

Cement Manufacturing - Dry

Cement Manufacturing - Wet

Coal Cleaning-Thrml Dryer; Fluidized Bed

Comm./lnst. Incinerators

External Combustion Boilers, Solid Waste

Fluid Cat Cracking Units; Cracking Unit

Fuel Fired Equip; Process Htrs; Pro Gas

Glass Manufacturing - Container

Glass Manufacturing - Flat

Glass Manufacturing - Pressed

ICI Boilers - Coal/Cyclone

ICI Boilers - Coal/FBC

ICI Boilers - Coal/Stoker

ICI Boilers - Coal/Wall

ICI Boilers - Gas

ICI Boilers - Natural Gas

ICI Boilers - Process Gas

ICI Boilers - Residual Oil

Indust. Incinerators

In-Proc;Process Gas;Coke Oven/Blast Furn

Recommended

Number of
Sources

15

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(A)
2.427

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(B)
1,551

Size of correction
in NOx emission
reductions
(tons/03 season)
(A-B)
875

14

1.199

0

1,199

36

3.932

6,586

-2,654

20

3.672

2,234

1,438

7

1.294

1,120

174

1

50

50

0

2

137

0

137

6

472

0

472

6

607

52

556

2

143

143

0

34

2,759

678

2,081

23

10,241

6,024

4,217

8

684

402

282

8

2,987

1,840

1,147

3

233

180

53

45

4,688

2,938

1,750

54

12,041

7,996

4,045

10

1,266

910

356

84

7,578

3,452

4,126

36

3,868

1,229

2,639

2

199

82

117

9

586

124

461

3

299

0

299

16


-------
Number of

Source Group	Sources

In-Process; Bituminous Coal; Cement Kiln	2

Iron & Steel - In-Process Coal Combustion	4

Iron & Steel Mills - Reheating	2

Municipal Waste Combustors	55

Nitric Acid Manufacturing	7

Petroleum Refinery Gas-Fired Process Heaters	28

Taconite Iron Ore - Induration - Coal or Gas	10

Utility Boiler - Coal/Wall	5

Utility Boiler - Oil-Gas/Tangential	2

Utility Boiler - Oil-Gas/Wall	4

547

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(A)
290

Recommended
NOx emissions
reduced from
controls in CoST
(tons/03 season)
(B)
295

Size of correction
in NOx emission
reductions
(tons/03 season)
(A-B)
-5

419

0

419

156

156

0

1.591

876

715

687

82

605

2.025

0

2,025

829

451

379

555

555

0

526

526

0

1.645

1,524

121

70,082

42,054

28,028


-------
Attachment 3 - NOx Emission Reductions by 3-Digit NAICS Code for Sources in the >100 Ton per Year Reduction Group







Recommended







NOx emissions

NOx emissions

Size of correction





reduced from

reduced from

in NOx emission





controls in CoST

controls in CoST

reductions



Number of

(tons/03 season)

(tons/03 season)

(tons/03 season)

3-Diait NAICS Code

Sources

(A)

(B)

(A-B)

211 Oil and Gas Extraction

1

46

30

16

212 Mining (except Oil and Gas)

11

879

500

379

221 Utilities

10

1.186

853

333

311 Food Mfg

12

1.181

815

366

312 Beverage and Tobacco Product Mfg

7

761

761

0

322 Paper Mfg

70

11.616

7,968

3,648

324 Petroleum and Coal Products Mfg

49

3.942

239

3,703

325 Chemical Mfg

132

19.689

10,753

8,937

3272 Glass and Glass Product Mfg

64

13,588

7,047

6,540

3273 Cement and Concrete Product Mfg

64

9,113

10,183

-1,070

3274 Lime & Gypsum Product Mfg

1

75

52

22

331 Primary Metal Mfg

50

4,908

1,837

3,070

333 Machinery Mfg

1

57

35

21

336 Transportation Equipment Mfg

2

148

103

46

424 Merchant Wholesalers, Nondurable Goods

2

160

0

160

531 Real Estate

1

72

0

72

562 Waste Mgmt and Remediation Services

65

2,366

843

1,523

611 Educational Services

5

295

34

261



547

70,082

42,054

28,028

18


-------
Attachment 4 - NOx Emission Reductions by 3-Digit NAICS Code for Sources in the > 100 Ton per Year Reduction Group







Recommended







NOx emissions

NOx emissions

Size of correction





reduced from

reduced from

in NOx emission





controls in CoST

controls in CoST

reductions



Number of

(tons/03 season)

(tons/03 season)

(tons/03 season)

Recommended Change to CoST Control

Sources

(A)

(B)

(A-B)

Already Controlled

138

12.973

0

12,973

Already Controlled by Glass CD

12

1.034

0

1,034

Already Controlled By Refinery CD

52

4.300

0

4,300

Control Technically or Economically Infeasible

18

1.618

0

1,618

Fuel Switch Already Occurred

4

2.370

0

2,370

Low NOx Burner

7

629

500

129

Low NOx Burner and Flue Gas Recirculation

88

8.792

6,022

2,769

Low NOx Burner and SCR

44

7.996

7,996

0

Low NOx Burner and SNCR

41

5,895

10,236

-4,341

Non-Selective Catalytic Reduction

1

82

82

0

Oxygen Enriched Air Staging

47

12,077

7,104

4,973

Selective Catalytic Reduction (SCR)

27

6,088

6,088

0

Selective Non-Catalytic Reduction (SNCR)

62

5,109

4,026

1,083

Source Already Shutdown

6

1,120

0

1,120



547

70,082

42,054

28,028

19


-------
Attachment 5 - NOx Emission Reductions by State for Sources in the 25 to 100 Ton per Year Reduction Group

Recommended





NOx emissions

NOx emissions

Size of correction





reduced from

reduced from

in NOx emission





controls in CoST

controls in CoST

reductions



Number of

(tons/03 season)

(tons/03 season)

(tons/03 season)

State

Sources

(A)

(B)

(A-B)

Alabama

38

641

517

123

Arkansas

14

277

203

74

Delaware

5

73

58

15

Florida

27

532

399

133

Illinois

91

1.519

845

675

Indiana

44

894

580

314

Iowa

19

422

309

113

Kansas

31

562

421

140

Kentucky

33

619

407

212

Louisiana

101

2.046

1.467

579

Maryland

18

353

209

144

Michigan

67

1.149

844

304

Mississippi

22

366

343

23

Missouri

13

224

179

45

New Jersey

7

72

11

61

New York

41

685

59

625

Ohio

86

1.476

1,075

402

Oklahoma

40

749

669

81

Pennsylvania

79

1.359

423

936

Tennessee

42

742

514

228

Texas

374

6,444

3,311

3,133

Virginia

30

450

350

100

West Virginia

21

421

334

87

Wisconsin

37

697

471

226



1280

22,774

14,000

8,774

20


-------
Attachment 6 - NOx Emission Reductions by Source Group for Sources in the 25 to 100 Ton per Year Reduction Group







Recommended







NOx emissions

NOx emissions

Size of correction





reduced from

reduced from

in NOx emission





controls in CoST

controls in CoST

reductions



Number of

(tons/03 season)

(tons/03 season)

(tons/03 season)

Source Group

Sources

(A)

(B)

(A-B)

Ammonia - NG-Fired Reformers2

7

200

155

45

Cement Kilns

4

93

0

93

Cement Manufacturing - Wet

2

60

0

60

Ceramic Clay Mfg; Drying

4

29

0

29

Coal Cleaning-Thrml Dryer; Fluidized Bed

10

188

188

0

Comm./lnst. Incinerators

4

47

47

0

Ext Comb Boilers, Elec Gen, Nat Gas (2)

1

28

28

0

Ext Comb Boilers, Elec Gen, Sub/Bit Coal (3)

14

158

158

0

Fbrglass Mfg; Txtle-Type Fbr; Recup Furn

2

9

9

0

Fluid Cat Cracking Units; Cracking Unit

21

393

0

393

Fuel Fired Equip; Furnaces; Natural Gas

3

18

18

0

Fuel Fired Equip; Process Htrs; Pro Gas

7

86

86

0

Gas Turbines - Natural Gas

438

7,193

5,749

1,444

Glass Manufacturing - Flat

8

190

0

190

ICI Boilers - Coal/FBC

1

35

22

13

ICI Boilers - Coal/Stoker

133

2,502

1,629

873

ICI Boilers - Coal/Wall

11

246

246

0

ICI Boilers - Distillate Oil

4

75

0

75

ICI Boilers - Gas

26

601

0

601

ICI Boilers - Natural Gas

350

6,814

3,705

3,109

ICI Boilers - Oil

2

41

0

41

ICI Boilers - Process Gas

31

609

0

609

ICI Boilers - Residual Oil

28

484

437

47

21


-------
Number of

Source Group	Sources

Indust. Incinerators	21

In-Proc;Process Gas;Coke Oven/Blast Furn	4

Iron & Steel - In-Process Comb - Coal	1

Iron & Steel Mills - Reheating	32

Municipal Waste Combustors	25

Nitric Acid Manufacturing	14

Petroleum Refinery Gas-Fired Process Heaters	30

Solid Waste Disp;Gov;Other lncin;Sludge	1

Space Heaters - Natural Gas	2

Steel Foundries; Heat Treating Furn	7

Surf Coat Oper;Coating Oven Htr;Nat Gas	2

Utility Boiler - Coal/Wall	2

Utility Boiler - Coal/Wall2	1

Utility Boiler - Oil-Gas/Tangential	8

Utility Boiler - Oil-Gas/Wall	19

1280

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(A)
230

Recommended
NOx emissions
reduced from
controls in CoST
(tons/03 season)
(B)
118

Size of correction
in NOx emission
reductions
(tons/03 season)
(A-B)
113

33

8

25

19

0

19

481

481

0

472

228

243

363

289

74

456

0

456

6

6

0

17

13

4

122

122

0

11

0

11

48

48

0

13

13

0

99

62

37

307

137

170

22,774

14,000

8,774


-------
Attachment 7 - NOx Emission Reductions by 3-Digit NAICS Code for Sources in the 25 to 100 Ton per Year Reduction Group

3-Diqit NAICS Code

211	Oil and Gas Extraction

212	Mining (except Oil and Gas)

213	Support Activities for Mining
221 Utilities

311	Food Manufacturing

312	Beverage and Tobacco Products

313	Textile Mills

314	Textile Product Mills

316 Leather and Allied Product Manufacturing

321	Wood Product Manufacturing

322	Paper Manufacturing

324	Petroleum and Coal Products

325	Chemical Manufacturing

326	Plastics and Rubber Products

327	Nonmetallic Mineral Product Manufacturing

331	Primary Metal Manufacturing

332	Fabricated Metal Product Manufacturing

333	Machinery Manufacturing

334	Computer and Electronic Products

336	Transportation Equipment Manufacturing

337	Furniture and Related Products
447 Gasoline Stations

454 Nonstore Retailers

Recommended

NOx emissions NOx emissions	Size of correction

reduced from reduced from	in NOx emission

controls in CoST controls in CoST	reductions

Number of (tons/03 season) (tons/03 season)	(tons/03 season)

Sources {A> (B)	{A-Bl

146

2.674

2,573

100

12

247

227

20

1

20

20

0

96

1.575

1,035

540

46

715

450

266

9

151

91

60

1

24

15

9

1

12

7

4

1

10

7

3

6

100

56

44

79

1,662

1,028

634

115

2,083

527

1,556

332

6,480

3,218

3,262

13

206

142

65

24

417

32

385

87

1,380

1,094

285

4

80

46

33

2

20

14

6

1

9

9

0

13

261

192

69

2

18

18

0

1

7

0

7

1

9

0

9

23


-------
Number of

3-Diqit NAICS Code	Sources

482 Rail Transportation	3

486 Pipeline Transportation	156

488 Support Activities for Transportation	1

531 Real Estate	8

541 Professional Services	6

561	Administrative and Support Services	1

562	Waste Mgmt and Remediation Services	21
611 Educational Services	62
622 Hospitals	7
713 Amusement, Gambling, and Recreation	2
721 Accommodation	2

922	Justice, Public Order, and Safety Activities	4

923	Administration of Human Resources	1
928 National Security and International Affairs	13

1280

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(A)
37

Recommended
NOx emissions
reduced from
controls in CoST
(tons/03 season)
(B)
23

Size of correction
in NOx emission
reductions
(tons/03 season)
(A-B)
14

2.551

2,034

517

18

12

6

147

0

147

81

77

4

8

0

8

376

184

192

963

617

346

116

36

80

49

45

4

25

10

15

29

17

12

12

8

4

201

135

66

22,774

14,000

8,774


-------
Attachment 8 - NOx Emission Reductions by 3-Digit NAICS Code for Sources in the 25 to 100 Ton per Year Reduction Group

Recommended Change to CoST Control

Already Controlled

Already Controlled by Refinery CD

Low NOx Burner

Low NOx Burner and Flue Gas Recirculation
Low NOx Burner and SCR
Low NOx Burner and SNCR
Natural Gas Reburn
Non-Selective Catalytic Reduction
Questions About Feasibility
Questions About Feasibility - Cement
Questions About Feasibility - Ceramic Clay Mfg
Questions about Feasibility - Coating Ovens
Questions about Feasibility - Distillate Oil
Questions About Feasibility - Glass
Questions about Feasibility - Process Gas
Selective Catalytic Reduction
Selective Non-Catalytic Reduction

Recommended

Number of
Sources

207

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(A)
3.380

NOx emissions
reduced from
controls in CoST
(tons/03 season)
(B)
0

Size of correction
in NOx emission
reductions
(tons/03 season)
(A-B)
3,380

40

704

0

704

362

6.087

6,087

0

361

6.726

4,491

2,235

11

246

246

0

24

437

437

0

1

28

28

0

12

289

289

0

1

22

0

22

6

154

0

154

4

29

0

29

2

11

0

11

6

116

0

116

7

167

0

167

50

1,070

0

1,070

8

216

216

0

178

3,094

2,207

886

1280

22,774

14,000

8,774

25


-------
From:

Subject:

Date:

To:

US EPA OAQPS
SRA International, Inc.

Summary of State NOx Regulations for Selected Stationary Sources
September 30, 2014

SRA compiled a summary of state/local NOx emission control regulations pertaining six categories of
nonEGUs:

•	Cement kilns

•	Industrial/Commercial/Institutional (ICI) Boilers - Coal-fired

•	ICI Boilers - Gas-fired

•	ICI Boilers - Oil-fired

•	Gas Turbines

•	Internal Combustion (IC) Engines

The analysis included 27 states in the eastern luo-ihuds of ihe I S I'or each of iliese states and source
categories, we identified state-specific sub-calegones (c g fuel l\ pe or si/.e ihreshold), the NOx emission
limit or control requirement, averaging time lor ihe emission liniil. geographic applicability within the
state, testing/monitoring requirements, and rule cilalion This mformalion is contained in the attached
spreadsheet (Draft State NOx RACT I.mills 2d 14 i>4 <)| \ls\).

Attachment 1 is an overall summaiy of ihe relali\e sliingcncv of the NOx requirements by geographic
area and source cak-ijoiy We also prepared a 2-page summary for each of the six categories to concisely
compare state NO\ emission limns or conliol requirements. These are shown in Attachments 2 to 7, along
withnoles highlighting the major differences Ivlweenthe state regulations.

Please lei us know should \ou ha\e questions or comments about any of the data presented in this
memorandum

1


-------
Attachment 1 - Relative Stringency of NOx Requirements

Source Category

States/Areas with
Most Stringent Regulations

States/Areas with
Less Stringent Regulations

States with
No Regulations or Sources

Cement Kilns1

States: IL, MD, NY, PA, TX
Areas: Ellis County, TX

States: AL (NOx SIP area), IN, KY,
MO, Ml, OH, SC, TN, VA, WV

States: AR, FL, GA, MS, OK
States with no cement kilns:
CT, DE, LA, MA, NC, NJ, Wl

Coal-fired ICI Boilers2

States: NY

Areas: Chicago, St. Louis (IL portion),
Baton Rouge, Houston-Galveston (coke-
fired), Milwaukee,

States: FL, GA, IN, MA, MD, Ml, PA,
TN, VA

Areas: Chicago, St. Louis (MO
portion), Baton Rouge, Charlotte,
Cleveland

States: AL, AR, KY, MS, OK, SC, TX
(except Houston-Galveston) WV
NE States with no coal-fired ICI
boilers: CT, DE, NJ

Gas-fired ICI Boilers

States: NJ, NY, PA

Areas: Chicago, St. Louis (IL portion),
Baton Rouge, Beaumont-Port Arthur,
Cleveland, Dallas, Houston, Milwaukee

States: CT, DE, FL,GA, MA, MD, Ml,
MO, TN, VA

Areas: Clark/Floyd Counties, St.
Louis (MO portion), Charlotte

States: AL, AR, KY, MS, OK, SC, WV

Oil-fired ICI Boilers

States: NJ, NY, PA

Areas: Chicago, St. Louis (IL portion),
Baton Rouge, Cleveland, Dallas, Houston,
Milwaukee

States: CT, DE, FL, GA, MA, MD,
Ml, TN, VA

Areas: Clark/Floyd Counties, St.
Louis (MO portion), Charlotte

States: AL, AR, KY, MS, OK, SC, WV

Gas Turbines

States: NJ

Areas: GA 45-county area, Dallas,
Houston, Milwaukee

States: CT, DE, FL, LA, MA, MD,
NY, PA, TN, VA
Areas: Chicago, St. Louis (IL
portion), St. Louis (MO portion),
Charlotte, Cleveland,

States: AL, AR, IN, KY, Ml, MS, OK,
SC, WV

IC Engines > about 500 hp

States: MD, NJ, NY

Areas: Chicago, St. Louis (IL portion),
Dallas, Houston

States: CT, DE, MA, Ml, PA, TN, VA
Areas: Baton Rouge, St. Louis (MO
portion), Charlotte, Cleveland,
Milwaukee

States: AL, AR, IN, KY, MS, OK, SC,
WV

1)	Cement kiln emission limits imposed by recent EPA enforcement settlements tend to be more stringent than the emission control
requirements in state rules.

2)	CT, DE and NJ have no active coal-fired boilers, so the stringency of their regulations for coal-fired ICI boilers is difficult to evaluate

2


-------
Attachment 2 - Cement Kilns



NOx Limit (lbs/ton clinker)

State

Long Dry

Long Wet

Pre-heater

Pre-calciner

AL

Ozone season: low-NOx burners, mid-kiln system firing, or approved ACT

AR

No Limits

No Limits

No Limits

No Limits

CT

No Cement Kilns in State

DE

No Cement Kilns in State

FL

No Limits

No Limits

No Limits

No Limits

GA

No Limits

No Limits

No Limits

No Limits

IL

5.1

5.1

3.8

2.8

IN

6.0

5.1

3.8

2.8

IN

(Clark/Floyd)

10.8 (op day)/
6 (30 day)

No Limits

5.9 (op day)/
4.4 (30 day)

No Limits

KY

6.6

6.6

6.6

6.6

LA

No Cement Kilns in State

MA

No Cement Kilns in State

MD

5.1

6.0

2.8

2.8

Ml

6.0

5.1

3.8

2.8

MO

6.0

6.8

4.1

2.7

MS

No Limits

No Limits

No Limits

No Limits

NC

No Cement Kilns in State

NJ

No Cement Kilns in State

NY

Case-by-case RACT Determination

OH

Ozone season: low-NOx burners, mid-kiln system firing, or approved ACT

OK

No Limits

No Limits

No Limits

No Limits

PA

3.44*

3.88*

2.36*

2.36*

SC

Ozone season:

ow-NOx burners, mid-kiln system firing, or approved ACT

TN

Ozone season: low-NOx burners, mid-kiln system firing, or approved ACT

TX

5.1

4

3.8

2.8

TX

(Ellis County)

No Limits

3.4

No Limits

1.7

VA

Case-by-case RACT Determination

Wl

No Cement Kilns in State

WV

Ozone season: low-NOx burners, mid-kiln system firing, or approved ACT

ACT = Alternative Control Technology

* Pennsylvania has proposed "RACT 2" presumptive RACT limits

3


-------
Observations Regarding State NOx Rules for Cement Kilns:

~	Geographic Applicability

•	All NOx SIP Call states with cement kilns have NOx rules in place

•	Since only portions of Alabama, Michigan, and Missouri were affected by NOx SIP Call, the
NOx rules only apply in the affected counties.

•	States not included in the NOx SIP Call do not have NOx RACT for cement kilns, except for
Texas. The Texas NOx requirements only apply in in Bexar, Comal. Ellis. Hays, and McLennan
Counties.

~	Form of NOx Limitation or Control Requirement

•	A few states express the requirement as "at least one of the follow mil: low -\()\ burners, mid-kiln
system firing, alternative control techniques or reasonably available control technology approved
by the Director and the EPA as achieving at leasl I lie same emissions decreases as with low-NOX
burners or mid-kiln system firing."

•	A few states specify presumptive emission li in i is in lemis of pounds ol' \()\ per ton of clinker.

•	Three states do not set presumptive emission limits but rather require facilities to submit a case-
by-case RACT determination. Pennsylvania has a proposed regulation that will specify
presumptive RACT limits: current rules require sources to hold 1 trading allowance per ton of
NOx calculated by multiplying ions clinker by the presumptive NOx limit.

~	Stringency of NOx Limitation or ( onliol Requirement

•	For states requiring "low-NOX burners, mid-kiln system firing, or ACT", it is generally assumed
that this will result in a 30% ivduclion from uncontrolled levels.

•	For states with numerical emission limits, the limits generally represent a 20 - 40 % reduction
from uncontrolled levels, depending on the type of kiln.

•	le\as has \eiy stringent limils for kilns in Ellis County.

•	Penns\ l\ ama lias proposed presumptive RACT emission limitations in April 2014 that are more
slnngenl lhan existing presumptive RACT limits in other states.

4


-------
Attachment 3 - Coal-fired Boilers





NOx Limit (Ibs/mmBtu)

State

Geographic Area

Boilers
50-100
mmBtu/hr

Boilers
100 - 250
mmBtu/hr

Boilers
>250
mmBtu/hr

AL

Statewide

No limits

No limits

No limits

AR

Statewide

No limits

No limits

No limits

CT

Statewide

0.29 to 0.43

0.29 to 0.43

0.29 to 0.43

DE

Statewide

LEA, Low NOx,
FGR

0.38 to 0.43

0.38 to 0.43

FL

Broward, Dade, Palm Beach
Counties

0.9

0.9

0.9

GA

45 county area

No limits

30 ppmvd @ 3%
02

0.7

IL

Chicago & St Louis areas

Tune-up

0.12 CFB
0.25 Other

0.12 CFB
0.18 Other

IN

Clark and Floyd Counties

No limits

0.4 to 0.5

0.4 to 0.5

KY

Statewide

No limits

No limits

No limits

LA

Baton Rouge 5 counties &
Region of Influence

0.2

0.1

0.1

MA

Statewide

0.43

0.33 to 0.45

0.33 to 0.45

MD

Select counties

No limits

0.38 to 1.0

0.38 to 1.0

Ml

Fine grid zone

No limits

No limits

0.4

MO

St Louis area

No limits

0.45 to 0.86

0.45 to 0.86

MS

Statewide

No limits

No limits

No limits

NC

Charlotte 6 county area

No limits

0.4 to 0.5

1.8

NJ

Statewide

0.43 to 1.0

0.38 to 1.0

0.38 to 1.0

NY

Statewide

No limits

0.08 to 0.20

0.08 to 0.20

OH

Cleveland 8 county area

0.3

0.3

0.3

OK

Statewide

No limits

No limits

No limits

PA

Statewide

0.45

0.45

0.20 to 0.35

SC

Statewide

No limits

No limits

NOx SIP Call

TN

5 Counties

Source specific
RACT

Source specific
RACT

Source specific
RACT

TX

Houston area

0.057
coke-fired

0.057
coke-fired

0.057
coke-fired

VA

Northern VA

No limits

0.38 to 1.0

0.38 to 1.0

Wl

Milwaukee 7 county area

0.10 to 0.25

0.10 to 0.25

0.10 to 0.20

WV

Statewide

No limits

No limits

No limits

5


-------
Observations Regarding State NOx Rules for Coal-fired Boilers:

~	Geographic Applicability

•	States in the OTR (CT, DE, MA, MD, NJ, NY, and PA) have NOx emission requirements that
apply statewide, not just in ozone nonattainment areas.

•	Six states (AL, AR, KY, MS, OK, and WV) do not have regulations limiting NOx emissions.

•	For the remaining states (FL, GA, IL, IN, KY, LA, MI, MO, NC, OH, TN, VA, WI), the NOx
emission control requirements only apply in ozone nonattainmenl areas

•	Texas only has emission limitations for coke-fired boilers in the I louslon-( ial\ cslon
nonattainment area.

~	Size Applicability

•	Most of the states do not have NOx emission requirements for boilers less than I no mmBtu/hour.

•	10 states do regulation boilers in the 50-100 mmBtu size range

~	Form of NOx Limitation or Control Requirement

•	Nearly all states express the NOx emission limits in terms of lbs/mmBtu.

•	A few states require either a case-bv-case RACT determination or specify specific types of
control equipment (e.g., low-NOx burners, flue gas recirculation).

~	Stringency of NOx Limitation or Control Requirement

•	Most states specify different emission limits for different types of boilers and firing types (e.g.,
dry bottom tangential-fired) vs. dry bottom wall-fired)

•	A few states in the Northeast have very few or no coal-fired ICI boilers, so the stringency of the
regulations in those states is difficult to evaluate. These states are CT, DE, NJ and MA.

•	For boilers greater than 100 mmBtu/hour. the LADCO/OTC1 Phase I recommended limits are in
the 0.2-0.3 lbs/mmBtu range (depending on boiler/firing configuration). The LADCO/OTC Phase
11 recommended limits arc in the 0.1 -0.2 lbs/mmBtu range. Four areas have limits that generally
meet ihe I.ADCO/OTC recommendations (Chicago, Baton Rouge, New York State, and
Milwaukee.

•	Texas has a \ ery stringent limit (0.057 lbs/mmBtu) for coke-fired boilers in the Houston-
Galveston area.

1 Evaluation of Control Options for Industrial, Commercial and Institutional (ICI) Boilers Technical Support
Document (TSD), March, 2011 prepared by the Lake Michigan Air Directors Consortium (LADCO) and the Ozone
Transport Commission (OTC).

6


-------
Attachment 4 - Gas-fired Boilers





NOx Limit (Ibs/mmBtu)

State

Geographic Area

Boilers
50-100
mmBtu/hr

Boilers
100 - 250
mmBtu/hr

Boilers
>250
mmBtu/hr

AL

Statewide

No Limits

No Limits

No Limits

AR

Statewide

No Limits

No Limits

No Limits

CT

Statewide

0.2 to 0.43

0.2 to 0.43

0.2 to 0.43

DE

Statewide

LEA, low NOx, FGR

0.2

0.2

FL

Broward, Dade, Palm Beach
Counties

0.2 to 0.5

0.2 to 0.5

0.2 to 0.5

GA

45 county area

30 ppmvd
@ 3% 02

30 ppmvd
@ 3% 02

0.2

IL

Chicago & St. Louis Areas

Tune-up

0.08

0.08

IN

Clark and Floyd Counties

No Limits

0.2

0.2

KY

Statewide

No Limits

No Limits

No Limits

LA

Baton Rouge 5 counties &
Region of Influence

0.1 to 0.2

0.1

0.1

MA

Statewide

0.1

0.2

0.2 to 0.28

MD

Select counties

Tune-up

0.2

0.2

Ml

Fine grid zone

No limits

Source specific
RACT

0.2

MO

St Louis area

No limits

0.2 to 0.5

0.2 to 0.5

MS

Statewide

No limits

No limits

No Limits

NC

Charlotte 6 county area

0.3

0.3

0.3

NJ

Statewide

0.1 to 0.5

0.1

0.1

NY

Statewide

0.05

0.06

0.08

OH

Cleveland 8 county area

0.1

0.1

0.1

OK

Statewide

No limits

No limits

No limits

PA

Statewide

0.08

0.08

0.08

SC

Statewide

No limits

No limits

No Limits

TN

5 Counties

Source specific
RACT

Source specific
RACT

Source specific
RACT

TX

Dallas and Houston areas

0.03 or
90% reduction

0.03 or
90% reduction

0.03 or
90% reduction

TX

Beaumont area

0.10

0.10

0.10

VA

Northern VA

0.2

0.2

0.2

Wl

Milwaukee 7 county area

No limits

0.08

0.08

WV

Statewide

No limits

No limits

No Limits

7


-------
Observations Regarding State NOx Rules for Gas-fired Boilers:

~	Geographic Applicability

•	States in the OTR (CT, DE, MA, MD, NJ, NY, and PA) have NOx emission requirements that
apply statewide, not just in ozone nonattainment areas.

•	Six states (AL, AR, KY, MS, OK, and WV) do not have regulations limiting NOx emissions.

•	For the remaining states (FL, GA, IL, IN, KY, LA, MI, MO, NC, OH, TN, TX,VA, WI), the NOx
emission control requirements only apply in ozone nonattainmenl areas

~	Size Applicability

•	About half of the states have NOx emission requirements for boilers less than I'll) mini itu/hour.
ranging from combustion tuning to emission limits as low as 0.05 Ilis mmlJlu

~	Form of NOx Limitation or Control Requirement

•	Nearly all states express the NOx emission limits in terms of lbs/mmBtu.

•	A few states require either a case-by-case RACT determination or specify specific types of
control equipment (e.g., low-NOx burners, flue gas recirculation).

~	Stringency of NOx Limitation or Control Requirement

•	The LADCO/OTC Phase I recommendations arc combustion tuning for boilers less than 100
mmBtu/hour. and either 0.1 lbs/mmBtu or 50% reduction for boilers greater than 100 mmBtu/hr.

•	The LADCO/OTC Phase II recommendations arc cither 0.05-0.1 lbs/mmBtu or 60% reduction.

•	New Jersey and New York have state-w ide limits that are consistent with the OTC/LADCO
Phase II recommendations. Pennsylvania has proposed state-wide limits that are consistent with
the OTC/LADCO Phase II recommendations.

•	Five areas (Chicago. Baton Rouge. Beaumont-Port Arthur, Cleveland, and Milwaukee) have
11hiiIs that arc consistent w ith the OTC/LADCO Phase II recommendations.

•	Dal las and I louslon ha\ c I he most stringent emission limitations - 0.02 lbs/mmBtu for greater
thai I'm ninililu hr umls

8


-------
Attachment 5 - Oil-fired Boilers





NOx Limit (Ibs/mmBtu



State

Geographic Area

Boilers
50-100
mmBtu/hr

Boilers
100 - 250
mmBtu/hr

Boilers
>250
mmBtu/hr

AL

Statewide

No limits

No limits

No limits

AR

Statewide

No limits

No limits

No limits

CT

Statewide

0.2 Distillate
0.25-0.43 Resid.

0.2 Distillate
0.25-0.43 Resid.

0.2 Distillate
0.25-0.43 Resid.

DE

Statewide

LEA, Low NOx, FGR

0.38 to 0.43

0.38 to 0.43

GA

45 county area

30 ppmvd

30 ppmvd

0.3

IL

Chicago & St Louis areas

Tune-up

0.1 Distillate
0.15 Resid.

0.1 Distillate
0.15 Resid.

IN

Clark and Floyd Counties

No limits

0.2 Distillate
0.3 Resid.

0.2 Distillate
0.3 Resid.

KY

Statewide

No limits

No limits

NOx SIP Call

LA

Baton Rouge

0.2

0.1

0.1

MA

Statewide

Tune-up

0.3 Distillate
0.4 Resid.

0.25 to 0.28

MD

Select counties

No limits

0.25

0.25

Ml

Fine grid zone

No limits

No limits

0.3 Distillate
0.4 Residual

MO

St Louis area

No limits

0.3

0.3

MS

Statewide

No limits

No limits

No limits

NC

Charlotte 6 county area

0.2

0.2

0.2

NJ

Statewide

Tune-up

0.1 Distillate
0.2 Resid.

0.1 Distillate
0.2 Resid.

NY

Statewide

0.08 to 0.2

0.15

0.15 to 0.2

OH

Cleveland 8 county area

0.12 Distillate
0.23 Resid.

0.12 Distillate
0.23 Resid.

0.12 Distillate
0.23 Resid.

OK

Statewide

New only

New only

New only

PA

Statewide

0.12 Distillate
0.20 Resid.

0.12 Distillate
0.20 Resid.

0.12 Distillate
0.20 Resid.

SC

Statewide

No limits

No limits

No limits

TN

5 Counties

Case-by-Case
RACT

Case-by-Case
RACT

Case-by-Case
RACT

TX

Dallas and Houston areas

No limits

~0.01

~0.01

VA

Northern VA

0.25 to 0.43

0.25 to 0.43

0.25 to 0.43

Wl

Milwaukee 7 county area

No limits

0.10 Distillate
0.15 Resid.

0.10 Distillate
0.15 Resid.

WV

Statewide

No limits

No limits

No limits

9


-------
Observations Regarding State NOx Rules for Oil-fired Boilers:

~	Geographic Applicability

•	States in the OTR (CT, DE, MA, MD, NJ, NY, and PA) have NOx emission requirements that
apply statewide, not just in ozone nonattainment areas.

•	Six states (AL, AR, MS, OK, SC, and WV) do not have regulations limiting NOx emissions.

•	For the remaining states (FL, GA, IL, IN, KY, LA, MI, MO, NC, OH, TN, TX, VA, WI), the
NOx emission control requirements only apply in ozone nonattai nine ill areas

~	Size Applicability

•	About half of the states have NOx emission requirements for boilers less iluin I < >() nunBtu/hour,
ranging from combustion tuning to emission limits as low as 0.08 lbs/inmBlu.

~	Form of NOx Limitation or Control Requirement

•	Nearly all states express the NOx emission limits in terms of lbs/mmBtu.

•	A few states require either a case-by-case RACT determination or specify specific types of
control equipment (e.g., low-NOx burners, flue gas recirculation).

~	Stringency of NOx Limitation or Control Requirement

•	The LADCO/OTC Phase I recommendations for distillate oil are combustion tuning for boilers
less than 100 nunBtu/hour. and either 0.1 lbs/mmBtu or 50% reduction for boilers greater than
100 mmBtu/hr. The LADCO/OTC Phase II recommendations for distillate oil are either 0.08-0.1
lbs/mmBtu or 60% reduction.

•	Only New Jersey has state-w ide limits that are consistent with the OTC/LADCO Phase II
recommendations for distillate oil.

•	Three areas (Chicago. Baton Rouge, and Milwaukee) have limits that are consistent with the
OTC I.ADCO Phase II recommendations for distillate oil.

•	The I A l)( () OTC Pluise I recommendations for residual oil are combustion tuning for boilers
less Hum Inn inmBtu hour, and either 0.2 lbs/mmBtu or 60% reduction for boilers greater than
100 mmBlu. hr. The LADCO/OTC Phase II recommendations for residual oil are either 0.2
lbs/mmBlu or 50-70% reduction.

•	New Jersey and New York have state-wide limits that are consistent with the OTC/LADCO
Phase II recommendations for residual oil. Pennsylvania has proposed state-wide limits that are
consistent with the OTC/LADCO Phase II recommendations for residual oil.

•	Four areas (Chicago, Baton Rouge, Charlotte, and Milwaukee) have limits that are consistent with
the OTC/LADCO Phase II recommendations for residual oil

•	Dallas and Houston have the most stringent emission limitations - 0.01 lbs/mmBtu for greater
that 100 mmBtu/hr units.

10


-------
Attachment 6 - Gas Turbines





NOx Limit (ppmvd @15% 02)

State

Geographic Area

Simple Cycle
>25 MW
Gas-fired

Simple Cycle
>25 MW
Oil-fired

Combined Cycle
> 25 MW
Gas-fired

Combined Cycle
>25 MW
Oil-fired

AL

Fine grid zone

No limits

No limits

No limits

No limits

AR

Statewide

No limits

No limits

No limits

No limits

CT

Statewide

55

258

(0.9 Ib/mmBtu)

55

258

(0.9 Ib/mmBtu)

DE

Statewide

42

88

42

88

GA

45 county area

6

6

6

6

IL

Chicago & St Louis
areas

42

96

42

96

IN

Statewide

No limits

No limits

No limits

No limits

KY

Statewide

No limits

No limits

No limits

No limits

LA

Baton Rouge 5
counties & Region
of Influence

54

(0.2 Ib/mmBtu)

86

(0.3 Ib/mmBtu)

54

(0.2 Ib/mmBtu)

86

(0.3 Ib/mmBtu)

MA

Statewide

65

100

42

65

MD

Select counties

42

65

42

65

Ml

Fine grid zone

No limits

No limits

No limits

No limits

MO

St Louis area

75

100

75

100

MS

Statewide

No limits

No limits

No limits

No limits

NC

Charlotte 6 county
area

75

95

75

95

NJ

Statewide

33

(2.2 Ib/MWh)

53

(3.0 Ib/MWh)

33

(2.2 Ib/MWh)

53

(3.0 Ib/MWh)

NY

Statewide

50

100

42

65

OH

Cleveland 8 county
area

42

96

42

96

OK

Statewide

No limits

No limits

No limits

No limits

PA

Statewide

42

75

42

75

SC

Statewide

No limits

No limits

No limits

No limits

TN

5 Counties

source specific
RACT

source specific
RACT

source specific
RACT

source specific
RACT

TX

Dallas and Houston
areas

9

(0.032 Ib/mmBtu)

9

(0.032 Ib/mmBtu)

9

(0.032 Ib/mmBtu)

9

(0.032 Ib/mmBtu)

VA

Northern VA

42

65/77

42

65/77

Wl

Milwaukee 7
county area

25 to 42

65 to 96

9

9

WV

Statewide

No limits

No limits

No limits

No limits

11


-------
12


-------
Observations Regarding State NOx Rules for Gas Turbines:

~	Geographic Applicability

•	States in the OTR (CT, DE, MA, MD, NJ, NY, and PA) have NOx emission requirements that
apply statewide, not just in ozone nonattainment areas.

•	Nine states (AL, AR, IN, KY, MI, MS, OK, SC, and WV) do not have regulations limiting NOx
emissions.

•	For the remaining states (GA, IL, LA, MO, NC, OH, TN, TX, VA. \\ I). llie NOx emission
control requirements only apply in ozone nonattainment areas.

~	Other Applicability Criteria

•	States use a variety size thresholds. For example. Ohio's rules diffeivniiale between units <3.5
MW and > 3.5 MW. Wisconsin has requiremenls lor lluve si/.e ranges: 10-25 MW. 25-50 MW,
and >50 MW.

•	State limits generally differ by type of fuel - yus or oil Wisconsin also includes limits for
biologically derived fuel.

•	Some states have different limits for simple-c\ cle and comlnned-c\ cle units. Other states have a
single limit that applies to both types of units.

~	Form of NOx Limitation or Control Rcquiremcnl

•	States do not specify specific types of control techniques, but rather set a numerical emission
limit.

•	Most states express limits in terms of "ppmv at 15% oxygen". Some states use lbs/mmBtu, and
the equivalent limits shown in the table above were calculated using based on Part 75 Eq-F5 and
F-factors. New Jersey's limits arc in terms of lbs/MHr.

~	Sli'ingenc\ of NOx Limitation or Control Requirement

•	Three areas ha\e \ei\ low limits compared to other states/areas: the 45 county area in Georgia,
Dallas and I louslon-(ial\eslon

13


-------
Attachment 7 - IC Engines Greater than ~500 hp





NOx Limit (g/hp-hr)

State

Geographic Area

Gas-fired,
Lean Burn

Gas-fired,
Rich Burn

Diesel

Dual Fuel

AL

Fine grid zone

No limits

No limits

No limits

No limits

AR

Statewide

No limits

No limits

No limits

No limits

CT

Statewide

2.5

2.5

8.0

8.0

DE

Statewide

Technology Stds.

Technology Stds.

Technology Stds.

Technology Stds.

GA

45 county area

?

?

?

?

IL

Chicago & St Louis
areas

210 ppmvd @

15% 02
(2.9 g/hp-hr)

150 ppmvd @

15% 02
(2.2 g/hp-hr)

660 ppmvd @

15% 02
(9.1 g/hp-hr)

660 ppmvd @

15% 02
(9.1 g/hp-hr)

IN

Statewide

No limits

No limits

No limits

No limits

KY

Statewide

No limits

No limits

No limits

No limits

LA

Baton Rouge 5
counties & ROI

4.0

2.0

?

?

MA

Statewide

3.0

1.5

9.0

9.0

MD

Select counties

150 ppmvd @

15% 02
(1.7 g/hp-hr)

110 ppmvd @

15% 02
(1.6 g/hp-hr)

175 ppmvd @

15% 02
(2.0 g/hp-hr)

125 ppmvd @

15% 02
(1.4 g/hp-hr)

Ml

Fine grid zone

3.0

1.5

2.3

1.5

MO

St Louis area

3.0 10.0

2.5 to 9.5

2.5-8.5

2.5-6.0

MS

Statewide

No limits

No limits

No limits

No limits

NC

Charlotte Area

2.5

2.5

8.0

8.0

NJ

Statewide

2.5

1.5

8.0

8.0

NY

Statewide

1.5

1.5

2.3

2.3

OH

Cleveland

3.0

3.0

3.0

3.0

OK

Statewide

No limits

No limits

No limits

No limits

PA

Statewide

3.0

2.0

8.0

8.0

SC

Statewide

No limits

No limits

No limits

No limits

TN

5 Counties

Source specific
RACT

Source specific
RACT

Source specific
RACT

Source specific
RACT

TX

Dallas and Houston
area

0.5

0.5

2.8 to 6.9

0.5

VA

Northern VA

Source specific
RACT

Source specific
RACT

Source specific
RACT

Source specific
RACT

Wl

Milwaukee 7
county area

3.0

3.0

3.0

3.0

WV

Statewide

No limits

No limits

No limits

No limits

14


-------
Observations Regarding State NOx Rules for IC Engines:

~	Geographic Applicability

•	States in the OTR (CT, DE, MA, MD, NJ, NY, and PA) have NOx emission requirements that
apply statewide, not just in ozone nonattainment areas.

•	Eight states (AL, AR, IN, KY, MS, OK, SC, and WV) do not have regulations limiting NOx
emissions.

•	For the remaining states (GA, IL, LA, MI, MO, NC, OH, TN, TX. Y.\. \YI). the NOx emission
control requirements only apply in ozone nonattainment areas.

~	Other Applicability Criteria

•	States use a variety size thresholds. For example, Louisiana's rules lui\ e separale linuis for IC
engines that are 150-300 hp, >300 hp, and > 15<>u hp New York uses -> 200 hp and 400 hp.
Delaware uses > 450 hp, while North Carolina uses foil hp

•	State limits generally differ by type of fuel - gas. oil. dual-fuel or landfill/digester gas.

•	A few states have different limits lean-burn and nch-lnun engines Other states have a single limit
that applies to both types of units.

~	Form of NOx Limitation or Control Requirement

•	Most states express limits in terms of "gram per brake horsepower hour".

•	Some states use "ppmvd @ 15% 02", and llie equivalent limits shown in the table above were
calculated using conversion factors from ppmv iin 15% 02 to g/hp-hr from EPA ACT, July 1993
EPA453-R-93-032.

•	Delaware specifies control technology standards rather than numerical emission limits.

~	Stringency of NOx Limitation or Control Requirement

•	\lar\ land. New Jerse\. New York and the Dallas/Houston areas of Texas have limits that are
more slnngenl lhan oilier slales/arcas.

15


-------