SECTION VI:

SUMMARY OF EMISSION CONTROLS AVAILABLE

FOR LARGE STATIONARY SOURCES OF NOx AND PM

Final Report

Prepared for:

Lee Alter

Western Governors’ Association

1515 Cleveland Place, Suite 200

Denver, CO 80202-5114

Prepared by:

Constance Senior, Brooke Shiley, Bradley Adams

Reaction Engineering International

77 West 200 South, Suite 210

Salt Lake City, UT 84101

Rui Afonso

Energy & Environmental Strategies

50 Old Faith Road

Shrewsbury, MA 01545

June 30, 2003

TABLE OF CONTENTS

1  INTRODUCTION  1

1.1  BACKGROUND  1

1.2  OBJECTIVES  1

1.3  DEFINITIONS AND METHODOLOGY  2

2  NO_X AND PM SOURCES IN THE WESTERN UNITED STATES  6

2.1  CHARACTERIZATION OF NO_X SOURCES  6

2.2  COMPARISON WITH OTHER DATABASES FOR NO_X CONTROL TECHNOLOGIES  13

2.3  TRENDS IN NO_X EMISSIONS AND CONTROLS FOR COAL-FIRED UTILITY BOILERS, 1995-2000.  16

2.4  CHARACTERIZATION OF SOURCES OF PARTICULATE MATTER (PM)  17

2.4  CHARACTERIZATION OF SOURCES OF PARTICULATE MATTER (PM)  18

2.5  COMPARISON WITH OTHER DATABASES FOR PM CONTROL TECHNOLOGIES  22

3  NO_X CONTROL TECHNOLOGIES  24

3.1  OVERVIEW  24

3.2  COAL-FIRED BOILERS  39

3.3  RECIPROCATING ENGINES  42

3.4  CEMENT KILNS  43

3.5  NATURAL GAS AND OIL FIRED BOILERS  43

3.6  TURBINES  44

4  PM CONTROL TECHNOLOGIES  46

4.1  OVERVIEW  46

4.2  PM CONTROL FOR COAL-FIRED BOILERS AND OTHER COMBUSTION SOURCES  47

4.3  OTHER DEVELOPMENTS  48

4.4  COSTS  49

5  MULTI-POLLUTANT CONTROL TECHNOLOGIES  51

5.1  PROPOSED MULTI-POLLUTANT EMISSION REGULATIONS FROM UTILITY BOILERS  51

5.2  MULTI-POLLUTANT CONTROL TECHNOLOGIES  52

6  SUMMARY AND RECOMMENDATIONS  56

6.1  NO_X AND PM SOURCES  56

6.2  CONTROLS FOR NO_X AND PM  57

6.3  WHAT’S ON THE HORIZON? WHAT TRENDS WILL INFLUENCE EMISSIONS AND CONTROL TECHNOLOGIES?  58

6.4  RECOMMENDATIONS FOR FUTURE WORK  59

7  REFERENCES  60

APPENDIX A: BREAKDOWN OF NO_X EMISSIONS BY STATE

APPENDIX B: BREAKDOWN OF PM EMISSIONS BY STATE

APPENDIX C: NO_X CONTROL TECHNOLOGY SUMMARIES

APPENDIX D: PM CONTROL TECHNOLOGY SUMMARIES

LIST OF ACRONYMS, ABBREVIATIONS AND SYMBOLS

ACFM    Actual cubic feet per minute

AFBC    Atmospheric fluidized bed combustor

BART  Best available retrofit technology

DLN    Dry Low NO_x

EIA    Energy Information Administration

EPA    United States Environmental Protection Agency

GCVTC  Grand Canyon Visibility Transport Commission

GCVTR  Grand Canyon Visibility Transport Region

FGR    Flue Gas Recirculation

Hg    Mercury

ICE    Internal Combustion Engine

LEC    Low Emission Combustion

LNB    Low-NO_x burner

MBtu    Millions of British Thermal Units

MTF    Market Trading Forum

NG    Natural Gas

NO_x    Nitrogen oxides

NSCR    Non-selective catalytic reduction

NSPS    New Source Performance Standard

O&M    Operating and Maintenance

OFA    Overfire air

PM    Particulate matter

PM10    Particulate matter less than 10 microns

PM2.5    Particulate matter less than 2.5 microns

SCC    Source Classification Code

SCR    Selective catalytic reduction

SIP    State Implementation Plan

SNCR    Selective non-catalytic reduction

SO₂    Sulfur dioxide

TPY    Tons per year

WGA    Western Governors’ Association

WRAP  Western Regional Air Partnership

ACKNOWLEDGEMENTS

We are grateful to Dr. Praveen Amar for providing help in locating the most recent sources of information on control technologies, as well as for external review of the report from the perspective of making it relevant to readership in the state governments/regional entities.

1  INTRODUCTION

1.1  Background

The Western Regional Air Partnership (WRAP) has undertaken a program to assess emissions control technologies and strategies for large stationary sources of NO_x and PM emissions in the western states region. The WRAP is a collaborative effort of tribal governments, state governments, and various federal agencies to implement the recommendations of the Grand Canyon Visibility Transport Commission (GCVTC) and to develop the technical and policy tools needed by western states and tribes to comply with the U.S. Environmental Protection Agency’s (EPA) Regional Haze Rule.

The WRAP established the Market Trading Forum (MTF), in large part, to develop and recommend emission control strategies for stationary sources of air pollution. A major focus of the MTF has been the establishment of regional emission milestones for sulfur dioxide (SO₂) and a regional backstop cap-and-trade program for SO₂ to be triggered if the milestones are not met voluntarily.

The MTF is also responsible for generating a report required in 40 CFR 51.309(d)(4)(v) of the Regional Haze Rule. The report must assess emission control technologies and strategies for stationary source NO_x and PM emissions and the degree of visibility impairment that would result from such strategies. It must also evaluate the need for NO_x and PM milestones to avoid any net emissions increase and to support possible multi-pollutant and multi-source control programs. Finally, this year several states must submit state implementation plans (SIPs) to EPA and must commit to a 2008 revision containing any necessary long-term strategies and Best Available Retrofit Technology (BART) requirements for stationary source NO_x and PM.

This project is essentially a starting point for addressing stationary source NO_x and PM emission sources over the next four years, at which point local and/or regional emission control program(s) may be implemented. Future work by the WRAP will investigate these issues further and will attempt more detailed cost estimates and emission reductions achievable in the WRAP region given the nature of its sources and existing controls.

1.2  Objectives

The main objectives of this project are to identify and briefly describe for large stationary sources in the western United States:

•  The universe of modern commercially-available or near-available stationary source NO_x and PM controls (either technologies or best management practices);

•  Trends in such controls;

•  Their approximate capital and operating costs, control efficiencies, and cost effectiveness;

•  Secondary environmental impacts, such as control of other air pollutants and generation of solid or hazardous waste;

•  Real-world experience at facilities implementing or testing such controls;

•  Future opportunities for improvements and demonstrations; and

•  Recommendations for future work.

1.3  Definitions and Methodology

The work plan for the project consisted of the following tasks:

Task 1. Inventory of Stationary Sources in the WRAP Region. This task involved a review of the 1996 WRAP stationary source emissions inventory (version 3, in MS Access format), as well as other recent and relevant databases to determine the number/type of stationary sources with emissions greater than 100 tons per year (TPY) and the type and performance of air pollution control devices installed on those sources. Two subsets were created for NO_x and PM emissions, respectively, based on the following criteria:

•  Sources (defined as emission units, or records, in the database) having annual emissions of the pollutant of interest greater than 100 TPY; and

•  Sources located in the thirteen-state region: AZ, CA, CO, ID, MT, ND, NM, NV, OR, SD, UT, WA, WY (See Figure 1).

Table 1 lists the fields extracted from the WRAP database.

Figure 1. Thirteen-state region considered in the technology assessment.

The source classification codes (SCCs) used to categorize sources served as general guidelines for choosing the categories in Task 1. The similarities (or differences) in the control technologies applicable to specific SCCs were also factors in grouping sources. For example, a category called “Coal-fired boilers” was created containing emissions data from utility and industrial boilers (of different boiler types) burning coal because the same NO_x and PM control technologies can be applied to most of these sources. With this in mind, Table 2 gives the categories created for characterization of the WRAP emissions and a description of the WRAP categories (i.e., SCC codes) used to define the categories in this report.

For electric utility point sources, additional databases were used to determine boiler capacity (MBtu/yr), enhance and update information on control technologies in place, and verify other source information. These databases were: EPA CEMs database for 1996 and 2001 [1], EPA E-GRID database for 1996 and 2000 [2] and the EIA-767 database for 1996 [3].

Table 1. WRAP database fields used in the technology assessment.

Note: Codes taken from the 1996 National Emission Trends (NET) PC Inventory File Format

The results of Task 1 are discussed in Section 2 of this report.

Table 2. List of Categories Used to Characterize Point Sources

Task 2. Survey and Documentation of Emission Control Technologies. In this task, we focused on the identification and compilation of control technologies for NO_x and PM (main focus) and for SO₂ and Hg (secondary focus). Sources identified in Task 1 that represented minor contributions to the emissions profile of the region, either due to their small number, uniqueness, or size, were considered in a more cursory fashion if their control technology options fell outside of the range of the more common/available technologies. This effort consisted mainly of literature reviews, on-line searches and personal (telephone) contacts and interviews.

The following information was collected on each technology or process:

•  Type and fundamentals of technology or process;

•  Projected performance;

•  Costs (capital and O&M or cost effectiveness in $/ton of pollutant removed) or cost projections;

•  Status of development and opportunities for or barriers to further development; and

•  Applicability to category (or categories) of WRAP sources identified in Task 1.

The results of this task are presented in Sections 3 and 4.

Task 3. Control Technology Analysis and Discussion. This task was the main focus of the project, in which a thorough evaluation and discussion of the many identified technologies was conducted. A summary containing the following information was created for each technology:

•  Process name

•  For each source category to which the technology was applicable, the following information was tabulated:

o  Total annual NO_x or PM emissions from sources greater than 100 TPY

o  Percentage NO_x and PM reduction  

o  Cost ($/ton or $/ACFM)

o  Development status

•  Detailed descriptions were prepared for the following:        

o  Process description

o  Achievable NO_x or PM reduction

o  Cost information

o  Development status

o  Practical considerations

o  Compatibility with other air pollution control technologies

o  Secondary environmental impacts

•  References

The results of this task are presented in Appendices C and D.

Task 4. Final Report. The draft version of the final report was submitted to WRAP on 25 April 2003. The final report was submitted on 30 June 2003.

2  NO_x AND PM SOURCES IN THE WESTERN UNITED STATES

2.1  Characterization of NO_x Sources

Table 3 gives the annual NO_x emissions in the GCVTR as well as in thirteen-state region for sources (defined as emission units, or records, in the WRAP database) exceeding 100 TPY. The cut-off of 100 TPY captures 84% of the stationary source NO_x emissions in the WRAP database for the thirteen-state region. Figure 2 shows the distribution of annual NO_x emissions (greater than 100 TPY) as a function of state.

The largest source category by far in the thirteen-state region is coal-fired boilers (69%); the top five categories (coal-fired boilers, internal combustion engines, cement kilns, turbines and oil and natural gas boilers) account for almost 90% of the NO_x emissions. Therefore, this report concentrates on control technologies applicable to these major process categories.

Figure 2. Annual NO_x emissions from sources with emissions greater than 100 TPY for the thirteen-state region.

The states with the largest NO_x emissions are AZ, CA, ND, NM, UT, and WY. Since all these states except ND are in the GCVTR, it is not surprising that emissions from the nine states in the GCVTR (AZ, CA, CO, ID, NM, NV, OR, UT, WY) account for 75% of the thirteen-state emissions greater than 100 TPY. Appendix A contains NO_x emissions by process category and by state.

Table 3. A comparison of annual emissions of NO_x from sources with emissions greater than 100 TPY between the thirteen-state region and the GCVTR.

With few exceptions, the distribution of NO_x sources is similar in the thirteen-state region as compared to the GCVTR. ICE’s (reciprocating engines and turbine) are predominantly in the GCVTR, while pulp and paper emissions are mostly outside the GCVTR. As a result of this similarity, the scope of this project was expanded to include additional WRAP states at minimal cost.

Figure 3. NO_x emissions from coal-fired utility boilers as a function of boiler type and coal rank for thirteen-state region from WRAP 1996 database.

The achievable NO_x emission rate depends on the fuel type. For coal-fired boilers, lower NO_x emission rates are obtained when firing subbituminous coal as compared to bituminous coal. Thus, it is useful to look at the distribution of coals in use in the thirteen-state region. Figure 3 shows the distribution of coals burned in utility boilers as a function of boiler type and coal rank. Most coal burned in the West is burned close to the mine; this distribution of coal rank reflects the native coals in the West.

For ICE’s, the application of NO_x control technology can depend on the type of fuel. More so than with utility boilers, the design and operation of the engine is often determined by the primary fuel. Most of the stationary ICE’s with annual emissions greater than 100 TPY burn natural gas, as shown in Figure 4.

Figure 4. NO_x emissions from Internal Combustion Engines as a function of engine type and fuel for thirteen-state region from WRAP 1996 database.

As long as a source category consists of primarily large sources, the cut-off of 100 TPY will include most of the NO_x emission sources. The 100-TPY cut-off captures 84% of the NO_x emissions in the WRAP database as a whole. However, certain source categories contain a very large number of small sources. For ICE’s (reciprocating engines and turbines) the 100-TPY cut-off only captures about 56% of the emissions as shown in Figures 5 and 6, although this is by far the second largest source category of stationary source NO_x emissions. Thus, NO_x control programs for sources in this category will require careful consideration of population attributes (e.g., controlling a large number of small sources).

Figure 5. Cumulative NO_x emissions from ICE’s in the thirteen-state region as a function of annual emission per source.

Figure 6. Cumulative NO_x emissions from ICE’s in the thirteen-state region as a function of number of sources (in order of decreasing annual emission per source.)

The 1996 WRAP database contains information on control technologies for the pollutants of interest. According to the 1996 data for sources greater than 100 TPY, few sources had NO_x controls, as shown in Table 4. Overall, just above 4% of the NO_x sources greater than 100 TPY in the WRAP 1996 database had installed controls. Coal-fired boilers were the most frequently controlled (15% of the units), followed by petrochemical processes (about 13% of the units). Note that control technologies listed in the right-hand column are as reported in the WRAP database. In a few cases, the description of the control technology does not seem correct (e.g., fabric filter or electrostatic precipitator) for NO_x control; this is a limitation of the data available and it is outside the scope of this program to determine the accuracy of the data in the WRAP database.

2.2  Comparison with Other Databases for NO_x Control Technologies

The level of control for coal-fired boilers in the WRAP database seemed low, even for 1996. Therefore, the 1996 WRAP database was compared with the data available for utility boilers in the 1996 CEMS and E-GRID databases. The EIA-767 database was also searched for NO_x control technologies. The E-GRID database should contain the information in the other two databases since it contains data from 24 different federal data sources, including EIA data and other EPA data. Only coal-fired utility boilers were included in this comparison, not all coal-fired boilers. However, only 3% of the WRAP NO_x emissions from coal-fired boilers in the thirteen-state region were from non-utility boilers.

It is worthwhile to take a closer look at utility boilers for two reasons. First, they are by far the largest source of NO_x emissions, accounting for 68% of the emissions from sources greater than 100 TPY. Second, the effectiveness of NO_x control technologies on boilers depends on the type of the boiler as well as on the fuel burned.

For this exercise, the EPA databases (CEMs and E-GRID) were queried to obtain information on capacity (MBtu per year) and control technologies. Data from 1996 was used in order to compare with the WRAP 1996 database. EPA and WRAP records were matched using ORIS Plant ID numbers and plant names. For matching records, control technologies not listed in the WRAP database were added, capacity (MBtu) entries were added, and NO_x emissions were replaced from the EPA databases.

A comparison of Tables 5 and 6, which contain, respectively, the WRAP data and the WRAP data augmented by the other databases, shows that the combination of the WRAP data and the EPA and EIA data suggests that 44% of the utility boilers had NO_x control (in 1996), as compared to only 12% when considering only the WRAP data by itself. The EPA databases probably undergo a more thorough QA/QC procedure than was used to create the WRAP database. Thus, the E-GRID and other federal databases might be expected to have more complete information.

The achievable NO_x emission rate depends on the boiler-fuel combination. The largest general class of utility boilers (in terms of number and capacity) is the dry bottom boiler. Dry bottom boilers can further be subdivided into wall-fired and tangential. Natural gas boilers emit less NO_x than coal-fired boilers per unit of fuel consumed. Of coal-fired boilers, tangential-fired units have the lowest emission rate and cyclones have the highest. The controls in the WRAP database are almost entirely low-NO_x burners or other combustion modifications. Figure 7 compares the range of NO_x emission rates for all boilers and fuels.

Application of low-NO_x burners and other combustion modifications can reduce NO_x emissions significantly; this can be seen in the large range of NO_x emission that is due, in part, to the use of NO_x controls on some of the boilers in each subset. Substantial NO_x reductions can also be achieved on coal-fired boilers just with combustion modifications.

Figure 7. Distribution of NO_x emission rates for utility boilers in the thirteen-state region, combination of WRAP and EPA/EIA databases for 1996.

Since 1996, low-NO_x burners have continued to improve; currently there are vendors who will guarantee NO_x emissions as low as 0.15 lb/MBtu from low-NO_x burners or low-NO_x firing systems. Furthermore, options have been developed for other combustion modifications, and SCR has begun to be applied to coal-fired boilers. Thus, the potential for NO_x control on coal-fired boilers is significantly better today than in 1996.

2.3  Trends in NO_x Emissions and Controls for Coal-Fired Utility Boilers, 1995-2000.

The most recent data available from the EPA databases for electric utility boilers are from 2000. In this section, we compare the 1996 data on NO_x emissions and controls discussed in the previous section with data from 2000.

Table 7 presents the data for 2000 derived from the EPA E-GRID and CEMS databases; this should be compared with Table 6 for 1996. The capacity of electric utility boilers increased by 37%, from 3,019,873,933 MBtu/yr in 1996 to 4,130,818,353 MBtu/yr in 2000, but the total NO_x emissions decreased by 7%. Figure 9 shows that the average annual emissions from dry bottom coal boilers (the largest category) decreased. Overall there was a decrease in emissions and an increase in the number of units that were controlled.

The number of sources increased, particularly the number of natural gas boilers, which increased from 47 to 82. The percent of natural gas boilers having NO_x controls decreased from 47% to 30%. During the time from 1996 to 2000, low-NO_x burners were added to natural gas units; there was also a small increase in SCR and SNCR on these types of boilers.

NO_x control on dry-bottom boilers increased from 47% to 71% from 1996 to 2000, resulting in a 9% decrease in total NO_x emissions from these boilers. The number of units with low-NO_x burners doubled. Overfire air (OFA) installations, though small in number, tripled. There were no SCR or SNCR installations on coal-fired boilers in 2000.

Thus, there was a modest reduction in NO_x emissions from electric utility boilers from 1996 to 2000, accompanied by a substantial increase in generating capacity. NO_x control increased, particularly on coal-fired boilers. The added NO_x control technologies were primarily low-NO_x burners and OFA.

2.4  Characterization of Sources of Particulate Matter (PM)

Table 8 gives the annual PM emissions for all PM sources in the thirteen-state region with emissions greater than 100 TPY. The cut-off of 100 TPY captures 60% of the PM emissions in the 1996 WRAP database for the thirteen-state region. With few exceptions, the distribution of PM sources is similar in the thirteen-state region as compared to the GCVTR. (Primary metal production emissions are mostly outside the GCVTR.) As a result of this similarity, the scope of this project was expanded to include additional WRAP states at minimal cost.

The largest source category (for those sources with emissions greater than 100 TPY) in the thirteen-state region is coal-fired boilers (40%); the top eight categories account for 92% of the PM emissions. Therefore, this report will focus on control technologies applicable to these process categories.

The state with the largest PM emissions is WY, followed by AZ, ID, and NM (Figure 9). Since all these states are in the GCVTR, it is not surprising that emissions from the nine states of the GCVTR (AZ, CA, CO, ID, NM, NV, OR, UT, WY) account for 83% of the total stationary source emissions greater than 100 TPY, as shown in Figure 10. Appendix B contains PM emissions by process category and by state.

Table 8. Annual Emissions of PM from Sources with Greater than 100 TPY.

.

Figure 8. Average Annual NO_x Emissions (greater than 100 TPY) from Electricity Generating Boilers: Comparison of 1996 and 2000 data from EPA Databases.

Figure 9. Annual PM Emissions from Sources with Emissions Greater than 100 TPY for the Thirteen-State Region.

Table 9 lists the control technologies in use in the 1996 WRAP database for particulate matter. 72% of coal-fired boilers, the largest category of emissions, had some form of PM control. Overall, though, only 38% of sources with emissions greater than 100 TPY had controls.

PM Control Technology

Centrifugal Collector (Cyclone)(2), Electrostatic Precipitator(35), Fabric Filter(12), Multiple Cyclone(4), Multiple Cyclone/Electrostatic Precipitator(2), Multiple Cyclone/Wet Scrubber(1), Wet Scrubber(8)

Centrifugal Collector (Cyclone)(2), Dust Suppression by Chemical Stabilizers or Wetting(5), Dust Suppression by Water Sprays(16), Fabric Filter(1), Water Curtain(1), Wet Scrubber(4)

Centrifugal Collector (Cyclone)(3), Sulfuric Acid Plant - Contact Process(2), Wet Scrubber(2)

Centrifugal Collector (Cyclone)(1), Wet Scrubber(1)

Dust Suppression by Water Sprays(2), Fabric Filter(1)

Alkalized Alumina(2), Dust Suppression by Water Sprays(1), Wet Scrubber(3), Misc.(1)

Centrifugal Collector (Cyclone)(1), Wet Scrubber(1)

None

None

Electrostatic Precipitator(4)

Centrifugal Collector (Cyclone)(1)

None

None

Electrostatic Precipitator(1)

None

None

None

None

Avg PM (TPY/Source)

523

288

258

238

512

235

298

444

2,456

276

230

233

419

160

537

267

175

176

357

Total PM (TPY)

46,010

24,499

10,836

5,718

5,631

4,697

4,476

3,555

2,456

1,379

1,150

932

838

641

537

535

525

176

114,590

Controlled Units

64

29

7

3

3

7

3

0

0

4

1

0

0

1

0

0

0

0

122

Number of Units

88

85

42

24

11

20

15

8

1

5

5

4

2

4

1

2

3

1

321

Coal-Fired Boilers

Mineral Processing

Petrochemical

Wood Boilers

Refinery Emissions

Primary Metal Production

Pulp and Paper

Smelting Operations

Miscellaneous

Oil/NG Boilers

Sugar Beet Processing

Cooling Tower

Turbines

Cement Kilns

Secondary Metal Production

Jet Engine Testing

Reciprocating Engines

Refinery Process Heaters

Total

2.5  Comparison with Other Databases for PM Control Technologies

The 1996 WRAP database was compared with the data available for utility boilers in the 1996 CEMS and E-GRID databases. The EIA-767 database was also searched for PM control technologies. The E-GRID database should contain the information in the other two databases since it contains data from 24 different federal data sources, including EIA data and other EPA data.

EPA and WRAP records for 1996 were matched using ORIS Plant ID numbers and plant names. For matching records, control technologies not listed in the WRAP database were added, capacity (MBtu) entries were added, and PM emissions were replaced from the EPA databases. The EIA-767 database reported PM emissions as lb PM/MBtu, from which we calculated PM emissions in tons per year.

Table 10. Sample PM Records from WRAP 1996 and EPA 1996 databases.

PM emissions data in the EPA databases do not agree with data in the WRAP database, suggesting that the data were obtained from different measurement and/or estimation methods. The differences, illustrated by a few sample records in Table 10, follow no general trend from plant to plant.

A comparison of Tables 11 and 12, which contain, respectively, the WRAP data and the WRAP data augmented by the other databases, shows that the combination of the WRAP data and the EPA and EIA data suggests that about 94% of the utility boilers had PM control (in 1996), as compared to only 53% when considering only the WRAP data by itself. The EPA databases probably undergo a more thorough QA/QC procedure than was used to create the WRAP database. Thus, the E-GRID and other federal databases might be expected to have more complete information.

PM Control Technology	Multiple Cyclone(3), Fabric Filter(8), Wet Scrubber(7), Electrostatic Precipitator(12)	Electrostatic Precipitator(4)	Electrostatic Precipitator(3)	None	Multiple Cyclone(1)	None
Average Emissions (TPY/Source)	615	309	234	190	381	102	554
PM Emissions (TPY)	36,889	1,235	702	571	381	102	39,880
Controlled Units	30	4	3	0	1	0	38
Number of Units	60	4	3	3	1	1	72
	Dry Bottom	NG Boiler	Cyclone	Stoker	Coal-fired AFBC	Wet Bottom	Total

3  NO_x CONTROL TECHNOLOGIES

3.1  Overview

Table 13. Annual NO_x emissions greater than 100 TPY from major source categories.

As discussed in Section 2, the NO_x emissions greater than 100 TPY in the thirteen-state region come predominantly from coal-fired boilers. We have concentrated on obtaining detailed information on NO_x control technologies for the top five categories, which account for 90% of the emission, although in some cases, where information was readily available, we have collected information for other source categories (refinery process heaters, glass melters, and wood-fired boilers). Table 13 shows that these source categories together account for 92% of the NO_x emissions greater than 100 TPY.

In this section, the information is organized in two formats. First, Table 14 lists all the technologies considered. For the most part, these are commercial technologies, in that vendors are offering these technologies. Not all technologies listed in Table 14 have demonstrated long-term operation, however. Table 14 gives the following information about each technology:

•  Name of the technology

•  Source categories to which the technology can be applied

•  Was a summary prepared? (Yes/No). If yes, technology summaries are contained in Appendix C.

Second, Tables 15 through 22 summarize the NO_x control options for major source categories for ease of comparison. More detailed information, particularly on the range of cost and NO_x control, is given in Appendix C. These tables contain the following information:

•  Name of Technology

•  Process Description

•  Applicability to units in the source category

•  Range of performance (NO_x removal efficiency)

•  Range of costs ($/ton of NO_x removed, levelized annual cost)

•  Commercial status

Table 14. NO_x Control Technologies.

Commercial	Y	Y	Y	Y	Y	Y
Cost, $/T	500-3000	200-1000	1000-2000	Not available	2000-10000	1300-3000
Performance	40-80% NOx reduction	30-60% NOx reduction	40-80% NOx reduction	70-99% NOx reduction claimed.	70-90+% NOx reduction	30-60% NOx reduction
Applicability	Most units, but could affect heat balance. Induced FGR requires pressure part changes.	Most boilers.	Most units. Furnace height may restrict some applications	Steam-hydrogen regeneration gas not practical for some boilers. Limited testing to date.	Most units. Space availability may constrain some options. High sulfur fuels more challenging	Most. Residence time and temperature characteristics are important.
Description	Recycle 15-25% of the flue gas to the windbox to reduce flame temperature. Can use eductors for induced FGR	Burners designed to produce lower NOx emissions – “staged” combustion	Form of “staged” combustion. Divert portion of the air from the windbox to OFA ports installed above the burners.	Add chemical reactor for NOx sorption, followed by regeneration.	Ammonia added upstream of catalytic reactor.	Inject ammonia-based reagent into upper furnace (1700-2000oF) to destroy NOx.
Technology	Flue Gas Recirculation (FGR)	Low-NOx Burners	Overfire Air (OFA)	SCONOX	SCR	SNCR

The formation of NO_x is a byproduct of the combustion of fossil fuels. Nitrogen contained in fuels such as coal and oil, as well as the harmless nitrogen in the air, will react with oxygen during combustion to form nitrogen oxides (NO_x). The degree to which this formation evolves is dependent on many factors, including both the combustion process itself and the properties of the particular fuel being burned. This explains why similar boilers firing different fuels or similar fuels burned in different boilers will yield different NO_x emissions.

As a result of these complex interactions in the formation of NO_x, an equally large number of approaches to minimize or reduce its emissions into the atmosphere have been and continue to be developed. A relatively simple way of understanding the many technologies available for NO_x emission control is to divide them into two major categories: (1) those that minimize the formation of NO_x during the combustion process (e.g., smaller quantities of NO_x are formed); and (2) those that reduce NO_x after the combustion process. It is common to refer to the first approach as “combustion modifications” whereas technologies in the second category are termed "post-combustion controls."

Within each of these categories, several technologies and variations of the same technology exist. Finally, combinations of some of these technologies are not only possible but often desirable as they may produce more effective NO_x control than the application of a stand-alone technology.

The following summaries describe the major technologies in each category.

3.2  Coal-Fired Boilers

Combustion modifications can vary from simple "tuning" or optimization efforts (similar to a "tune-up" of a car) to the deployment of dedicated technologies such as low-NO_x burners (LNB), Overfire air (OFA), or Reburn. All combustion modification approaches face a common challenge: that of striking a balance between NO_x reduction and fuel efficiency. The concern is exemplified by the typically higher carbon levels in the fly ash, which reflect lower efficiency (more fuel needed for the same electrical output) and which may contaminate the fly ash itself, possibly making it unsuitable for reutilization (e.g., cement and concrete production).

Combustion Optimization

Combustion optimization efforts can lead to reductions in NO_x emissions of 5%-15% or even higher in cases where a unit may be poorly "tuned." It is important to remember that optimization results are truly a function of the "pre-optimization" condition of the power plant or unit, and as such have limited opportunity for drastic emission reductions. Recent development of "intelligent controls" - software-based systems that "learn" to operate a unit and then maintain its performance during normal operation may go a long way towards keeping plants well-tuned as they age.

LNB’s and OFA

LNB’s and OFA represent practical approaches to minimizing the formation of NO_x during combustion. Simply, this is accomplished by controlling the quantities and the way in which fuel and air are introduced and mixed in the boiler (usually referred to as "fuel or air staging"). These technologies are the most prevalent in the power industry at present. For example, plants that have had to comply with Phase I of Title IV of the CAAA of 1990 have largely used these technologies for compliance. (Phase II of the Title IV has required the use of post-combustion technologies to meet more stringent requirements for both Group 1 and Group 2 boilers.) Competing manufacturers have proprietary designs, geared towards application in different boiler types, as well as reflecting their own design philosophies. LNB’s and OFA, which can be used separately or as a system, are capable of NO_x reductions of 40% - 60% from uncontrolled levels. Again, the type of boiler (e.g., dry vs. wet-bottom, wall- vs. tangential-fired, NSPS vs. pre-NSPS) and the type of fuel (e.g., bituminous vs. sub-bituminous) will influence the actual performance achieved. NO_x emission rates on the order of 0.15 lb/MBtu can be achieved with low NO_x burners under circumstances, particularly in dry-bottom boilers burning low-rank coals.

LNB’s/OFA have little or no impact on operating costs, other than those noted above. As such, the economics of these technologies are driven by capital/retrofit costs which typically range from $10-$40/kW, with the lower range reflecting easier "plug-in" application, whereas the higher costs are typically associated with more difficult and involved retrofits (e.g., where new controls or other systems may be replaced as part of the LNB retrofit).

From the standpoint of scheduling retrofits for existing units, LNB/OFA retrofit projects have "lead" times of 10-14 weeks and can require outages of 6-10 weeks, depending on factors such as scope of work, integration with other plant outage requirements, etc.

Reburn

Reburn, while generically included in the "Combustion Modification" category, is different from the other technologies in this group (LNBs/OFA) in that it "destroys" NO_x through chemically reducing conditions shortly after it is formed rather than minimizing its formation as discussed previously. From a practical standpoint, this is accomplished by introducing the reburn fuel (theoretically any fossil fuel can be used, but natural gas is the most common) into the boiler above the main burner region. Subsequently, this "fuel-rich" environment reacts with and "destroys" the NO_x formed in the main burners. This technology has been implemented in the U.S. and overseas, and while not as common as LNB/OFA, it is commercial at this time. Owing to stricter compatibility criteria, reburn is not as universal as LNB/OFA in its applicability to the overall boiler population. Specific criteria such as boiler size, availability of natural gas, type and quality of the main fuel are all important in determining the suitability of a unit for this technology. One important feature of reburn is its compatibility with cyclone boilers, for which the previously mentioned technologies are not particularly well suited. Cyclones boilers represent over 25,000 MW of capacity in the U.S.

Reburn performance has been shown to range from 35%-60% depending on such factors as reburn fuel type and quantity, initial NO_x level, boiler design, etc. Reburn can be thought of as a "dial-in" NO_x technology in that NO_x reductions are a function of the amount of reburn fuel. This feature may provide strategic value in compliance scenarios.

With respect to cost, systems using natural gas as the reburn fuel range from $15/kW to $30/kW whereas those using coal for reburn range from $30/kW to $60/kW. Operating costs are primarily driven by the fuel cost differential in the case of gas reburn, while for coal or oil reburn fuel preparation costs (pulverization and atomization, respectively) represent the dominating O&M costs. Countering these costs, particularly in the gas reburn case, are SO₂, particulates, and CO₂ co-benefits proportional to the fraction of gas used.

Project retrofit schedules for this technology are on the order of 15 to 20 weeks with 6 to 10 weeks of outage time likely.

Recently, reburn technology has evolved into several variations of the original approach. One of these is “Fuel Lean Gas Reburn" (FLGR) developed for specific applications where NO_x reductions of around 30%-40% may be required. FLGR uses less gas than conventional reburn (3%-7% vs 15%-20%), and its capital cost is less than $10/kW, making it a potentially effective option in specific applications.

SCR and SNCR

Readily available post-combustion NO_x controls are limited to selective non-catalytic reduction (SNCR) and selective catalytic reduction (SCR). They are fundamentally similar in that both use an ammonia-containing reagent to react with the NO_x produced in the boiler and convert it to nitrogen and water. SNCR accomplishes this at higher temperatures (1700ºF-2000ºF) in the upper furnace region of the boiler, while SCR operates at lower temperatures (about 600ºF) and hence needs a catalyst to produce the desired reaction between ammonia and NO_x.

While this difference between the two technologies may seem minor, it results in significant difference in performance and costs. This is because in the case of SNCR, the reaction occurs in a somewhat uncontrolled fashion (e.g., the existing upper furnace becomes the "makeshift" reactor, which is not what it was originally designed to be), while in the SCR case, a dedicated reactor and the reaction-promoting catalyst ensure a highly controlled, efficient reaction. In practice, this means that SNCR has lower capital costs (no need for a reactor/catalyst); higher operating costs (lower efficiency means that more reagent is needed to accomplish a given reduction in NO_x); and finally, has limited NO_x reduction capability (typically 30%-40%, with some cases achieving reductions in the 50% range). SCR, on the other hand, has higher capital costs but offers lower operating costs and the opportunity for very high NO_x reductions (up to 90%).

Capital costs range from $10 to $15/kW and $60 to $100/kW for SNCR and SCR, respectively. Operating costs are driven primarily by the consumption of the chemical reagent – usually urea for SNCR and ammonia for SCR, which in turn is dependent upon the efficiency of the process (usually referred to in terms of reagent utilization) as well as the initial NO_x level and the desired percent reduction. Two additional parameters important in the overall operating costs are (1) the potential contamination of fly ash by ammonia, making it unusable and (2) the life cycle of the catalyst due to fly ash.

Combined Approaches

In theory, most of these technologies can be used together. However, NO_x reductions are not necessarily additive, and more importantly, the “economics” of the combined technologies may or may not be cost-effective. Such analyses are highly site- and strategy-specific.

However, several such combinations of technology are considered attractive and have or are gaining acceptance. For example, the combination of LNB/OFA with either SCR or SNCR is more prevalent than the application of the post-combustion technologies alone. The economics of this approach are justified by the reduced chemical (SNCR) and capital costs (SCR – smaller reactor/catalyst) due to lower NO_x levels entering the SCR/SNCR system. Another example is the combination of Reburn with SNCR, driven by the synergisms between the two (similar location, temperatures in the boiler). This application may yield NO_x reductions of 60%-70% with capital costs in the $20-$30/kW range, but has a relatively high operating cost due to reagent and reburning fuel consumption.

3.3  Reciprocating Engines

Several control technologies are available for ICE’s, having a wide range of complexity, cost and performance.

Some in-cylinder methods offer low to moderate NO_x reductions at costs well below $1,000/ton. These include injection timing retard, and air/fuel ratio adjustment (with or without high-energy ignition). These methods are widely available, and NO_x performance will vary from one engine design to another. However, fuel efficiency can suffer as a result of these methods and emissions of products of incomplete combustion can increase.

Spark-ignited engines that can be retrofitted with Low-Emission Combustion (LEC) technology can potentially achieve significant NO_x reductions (80 to 90%). LEC technology can be expensive to retrofit on some engines, and it may not be available from all engine manufacturers. For large, low-speed engines, LEC technology is estimated to provide annual NO_x reductions of about 80% at under $1,000/ton under most conditions. LEC technology is estimated to be more cost effective on smaller, medium-speed engines (under $500/ton for annual control under most conditions). It is estimated to be somewhat more expensive for dual-fuel engines (annual control at a capacity factor of 65% is estimated to cost under $1,000/ton).

SCR is the only commercially available choice for post-combustion control of diesel and lean-burn spark-ignition engines. Experience in the U.S. with SCR on these engines is growing, especially for diesel engines. SCR has been applied to approximately 30 diesel engines and to an equivalent number of constant-load lean burn ICE’s. Experience with SCR on variable-load engines is limited. In analysis using data from case studies, it was estimated that SCR provides annual NO_x reductions of as high as 90% at a cost below $1,000/ton in all cases, except for very low capacity factors (~10%), and it provides seasonal reductions at a cost of under $1,000/ton for engines operating at high capacity factors (typically, 65% or greater).

Recent developments from the application of urea-SCR on mobile sources (diesel trucks) offer the possibility of reducing the size and capital cost of SCR systems for stationary ICE’s. This new technology, developed from efforts to apply SCR to mobile diesel engines, appears to make it possible to achieve much more cost-effective NO_x reduction on stationary ICE’s that operate for only a few hundreds of hours a year. NO_x reduction of about 75% is estimated to be possible for under $2,000/ton even for seasonal controls of some stationary ICE’s that operate only a few hundred hours each ozone season. Seasonal control at a cost of under $1,000/ton is estimated to be achievable for most applications with capacity factor greater than 45%.

3.4  Cement Kilns

As with other combustion systems, modifying the combustion process is one strategy for reducing NO_x in cement kilns. However, the quality of the clinker produced by the kiln can be affected by combustion modifications so these must be undertaken carefully.

Monitoring temperature and excess air in the combustion zone increases the efficiency of the cement-making process and can result in reduced NO_x emissions. Combustion modifications include staged combustion of air or fuel. Specifically designed low-NO_x burners are sometimes used. Even without low-NO_x burners, staging can be achieved by adding some of the fuel mid-kiln, as in mid-kiln injection of tires. Mid-kiln injection of fuel (most often tires) was in practice in twenty kilns in the U.S. in 2000.

Iron addition (CemStar process) has been used at about a dozen facilities in the U.S. This reduces the temperature needed in the kiln for formation of clinker and allows the combustion zone to operate cooler (and thus reducing NO_x).

Post-combustion (post-kiln) NO_x controls include SCR and SNCR. SCR has not been used on cement kilns in the U.S.; pilot studies have been conducted in Europe. SNCR technology requires a specific temperature window and residence time; these are not attainable in all cement kilns. SNCR can be applied to preheater/precalciner kilns. SNCR is widely practiced in Europe on cement kilns, but to date there have been only a handful of demonstrations in the U.S.

3.5  Natural Gas and Oil Fired Boilers

The menu of NO_x control options for gas and oil-fired boilers is essentially the same as for coal-fired boilers. One noted exception is the use of Flue Gas Recirculation (FGR), which is not effective in coal applications and hence, is not mentioned there.

While the control technologies are common to the coal-fired options, application issues require different considerations and analyses. Examples range from differences in the inherent NO_x formation amongst the fuels (thermal NO_x vs. “fuel”-NO_x), which dictate that combustion-based technologies are designed accordingly for each fuel, to the fact that gas produces no PM or SO₂/SO₃ and hence can afford some design changes from coal and oil applications. Equally important are the economics of the different fuels, which may favor different technology approaches.

In summary it can be said that the available menu of technologies is the same as for coal applications, but that (at least for gas), deployment of these technologies tends to be less constraining than for coal.

3.6  Turbines

There have been some important developments in gas turbine NO_x control technology, but well-established technologies continue to play an important role in reduction of NO_x. Dry Low NO_x (DLN), catalytic combustion, and some new post-combustion methods are making their way into the control technology market, while water or steam injection and SCR continue to be important technologies for reducing NO_x from gas turbines.

Many turbine manufacturers can convert or replace conventional combustors on existing turbines with DLN combustors. DLN combustion retrofits have been made possible by recent developments in gas turbine combustor technology. DLN technology offers the potential for substantial reduction of NO_x from turbines firing natural gas or other low-nitrogen fuels, as well as improved engine performance when compared to wet controls (water or steam injection). For turbines under about 15 MW in size, NO_x emissions of 25 ppm can be guaranteed for new turbines and emissions below 42 ppm can be guaranteed for retrofitted turbines. For large turbines (75 MW and higher in size), controlled NO_x emission levels of as low as 9 ppm have been guaranteed, even for retrofits.

DLN capital costs vary with the size of the turbine and the specifics of the particular turbine being retrofitted. The baseline NO_x level significantly affects the estimate of cost per ton of NO_x reduced. Using expected baseline NO_x emissions levels provided by the turbine manufacturers and retrofit costs expected to be typical of most applications, retrofit of DLN on industrial turbines (about 3 to 10 MW) originally equipped with conventional combustion control is estimated to provide NO_x reductions under $2,000/ton for annual controls with high capacity factors and at a higher cost for seasonal controls. For larger turbines (~75 MW), cost was estimated to be well below $1,000/ton for nearly all conditions.

Water injection and steam injection are two well-established technologies that can offer controlled NO_x emission levels below 42 ppm in many cases. Because water or steam injection technologies frequently have lower capital cost than DLN but higher variable costs, these technologies can be more attractive for peaking turbines or other turbines that operate infrequently. It was estimated that water injection installed on peaking units that operate 200 hours to 400 hours in the summer would reduce NO_x at a cost of about $2,500/ton to about $7,000/ton, depending upon the number of operating hours and the fuel used (gas or distillate oil).

SCR continues to be the most widely used post-combustion technology for gas turbines. Catalyst technology developments have made SCR viable over a wider temperature range. This makes SCR a viable control option in situations that were difficult in the past, such as simple-cycle turbines that may now benefit from high-temperature SCR and combined-cycle turbines with duct burners that may now benefit from low-temperature SCR.

The cost of NO_x reduction with SCR varies considerably according to application, turbine size, and the type of SCR technology that is appropriate for the application. As in the case of the DLN cost estimates, expected baseline NO_x emissions levels provided by the turbine manufacturers were used as a basis for cost calculations. Conventional SCR on a large (~75MW) combined-cycle turbine with high capacity factors was estimated to cost about $440/ton for annual controls and $870/ton for seasonal controls, for turbines equipped with conventional combustion technology (baseline NO_x emissions of 154 ppm). For turbines with lower baseline NO_x emissions (such as those equipped with DLN combustors having baseline NO_x emissions of 15 ppm), the cost per ton of additional NO_x removed was estimated to be greater, ranging from about $3,700/ton (annual control, high capacity factor) to over $13,000/ton (seasonal controls, low capacity factor). On smaller turbines (~5 MW), the cost of conventional SCR is estimated to be as low as $1,300/ton (with annual control and conventional combustion technology having baseline NO_x emissions of 142 ppm). Seasonal controls for smaller turbines are estimated at over $15,000/ton of NO_x removed at a low capacity factor (45%) with baseline NO_x emissions of 42 ppm.

For installations that may be better suited for high- or low-temperature SCR variants, such as simple-cycle turbines (high-temperature SCR) or combined-cycle turbines with limited space (low-temperature SCR), the cost of SCR is somewhat higher than for conventional SCR on a combined-cycle plant. A 75 MW turbine at a high capacity factor equipped with conventional combustion technology (baseline NO_x emissions of 154 ppm) can be controlled annually with high- or low-temperature SCR for about $550/ton and for about $1,200/ton seasonally. As with conventional SCR, turbines with lower baseline NO_x emissions (such as those equipped with DLN combustors) showed a higher cost per ton of NO_x reduction. The estimated cost of NO_x reduction for a 75 MW turbine with baseline NO_x emissions of 15 ppm ranges from $5,170/ton (annual controls, high capacity factor of 85%) to as high as $20,000/ton (seasonal controls, low capacity factor of 45%). On smaller turbines (~5MW), the cost for high- or low-temperature SCR is estimated to be as low as $2,000/ton with annual control and conventional combustion technology (baseline NO_x emissions of 142 ppm). Cost is estimated to range from $6,750/ton (annual controls, high capacity factor of 85%) to about $27,000/ton (seasonal controls, low capacity factor of 45%) with baseline NO_x emissions of 42 ppm.

Emerging combustion technologies (such as catalytic combustion) and post-combustion technologies (such as SCONOx) offer the potential for very low NO_x emission levels. Because there is much less experience with these technologies, available cost information is limited.

4  PM CONTROL TECHNOLOGIES

4.1  Overview

As discussed in Section 2, 8 source categories make up about 92% of the PM emissions and are summarized in Table 23. Detailed information on PM control technologies has been obtained for industrial processes that generate particulate matter. We have not provided cost information on fugitive emissions, however, since costs of fugitive dust control are highly variable and it is difficult to find an adequate metric for costs and then quantify them.

Table 23. PM emissions from top eight source categories.

Table 24 lists all the technologies considered. These are commercial technologies, in that vendors are offering these technologies with demonstrated operating experience in a wide range of applications. Table 24 gives the following information about each technology:

•  Name

•  Source categories to which the technology can be applied

•  Summary prepared? (Y/N)

Technology summaries are contained in Appendix D.

Table 24. PM Control Technologies

4.2  PM Control for Coal-Fired Boilers and Other Combustion Sources

Particulate matter is generated by a variety of physical and chemical processes. It is emitted to the atmosphere through combustion, industrial processes, fugitive emissions and natural sources. In combustion processes, the mineral matter (inorganic impurities) is converted to ash. The particles suspended in the flue gas are known as fly ash. Fly ash constitutes the primary particulate matter, which enters the particulate control device. Particulate matter is in general referred to as "PM", "PM10", "PM2.5" (particulate matter (PM) with an aerodynamic equivalent diameter of 10 microns or less and 2.5 microns or less, respectively).

Quantity and characteristics of the fly ash and particle size distribution depend on the mineral matter content of the fuel, combustion system, and operating conditions. Combustion technique mainly determines the particle size distribution in the fly ash and hence the final particulate emissions. Common combustion systems in pulverized coal firing include dry bottom, wall (front, opposed) and corner (tangential) burners and wet bottom furnaces. In dry bottom boilers, 10-20% of the ash is discharged as dry, bottom ash. In wet bottom boilers, 50-60% of the ash is discharged at the bottom of the boiler as slag. Stokers or grate-fired boilers are used to burn coal, wood and waste. The majority of the ash falls through the grate and is discharged as bottom ash. Mineral composition of the coal and the amount of carbon in the fly ash determine the quantity, resistivity and cohesivity of the fly ash.

PM emissions from other point source processes involve similar phenomena where particulate matter is carried with the flue gas, in suspension to the stack. Hence, the general technologies applicable to one source are typically suitable for the others as well. Factors such as type and quantity of PM, characteristics of the process gas (temperature, moisture, other contaminants) will have a major influence on the selection and design of the PM control technology.

Without getting into the details of the various technologies, the following four major types of particulate controls technologies are common for a variety of applications:

•  Wet scrubbers – scrubbers work on the principle of rapid mixing and impingement of the particulate with the liquid droplets and subsequent removal with the liquid waste. For particulate controls the “venturi scrubber” is an effective technology whose performance is directly related to the pressure loss across the venturi section of the scrubber. Venturi scrubbers are effective devices for particulate control. However, for higher collecting efficiencies and a wider range of particulate sizes, higher pressures are required. High-energy scrubbers refer to designs operating at pressure losses of 50-70 inches of water. Of course, higher pressure translates to higher energy consumption. Performance of scrubbers varies significantly across particle size range with as little as 50% capture for small (<2 microns) sizes to 99% for larger (>5 microns) sizes.

•  Electrostatic Precipitators (ESP) –ESP’s operate on the principle of electrophoresis, by imparting a charge to the particulates and collecting them on opposed charged plates. Dry vs. wet refers to whether the gas is water cooled and saturated prior to entering the charged plate area, or is collected dry on the plates. In gases with high moisture content, dry ESP’s are not suitable because the wet gas would severely limit the ability to collect the “sticky” particulates from the plates. The wet ESP technology is capable of very high removal efficiencies and is well-suited for the wet gas environments. Both types of ESP’s are capable of 99+% removals for particle sizes above 1 micron.

•  Fabric Filters – These are essentially “giant” vacuum cleaners. As in the case of the dry ESP, Fabric Filters are not well suited for wet gas applications. However FFs are extremely efficient in collecting PM including fine (submicron) size fractions.

•  Cyclones – Cyclones are devices that separate particulates from the gas stream through aerodynamic/centrifugal forces. However, the technology is only effective in removing larger size particulates (greater than about five microns).

4.3  Other Developments

While the technologies above represent the major available options for particulate control from point sources, it is relevant to note that advancements and innovative application of these technologies have and will continue to occur. Examples of these can vary from simple retrofits (e.g. new filter bag materials for Fabric Filters or newer spark control electronics on ESPs) to innovations including electrostatically- enhanced fabric filtration and hybrid concepts that combine attributes of various technologies.

The Electric Power Research Institute’s (EPRI) COHPAC process and the University of North Dakota Energy and Environmental Research Center’s Advanced Hybrid Particulate Collector (AHPC) are examples of hybrid particulate collectors. In COHPAC, an ESP is followed by a pulse-jet Fabric Filter either immediately following it or actually integrated into the original casing of the ESP (in the case of larger older ESP’s), where the FF acts as a “polishing” device significantly increasing the overall and fine particulate collection efficiency of the ESP alone. The AHPC technology can be described as an ESP with alternating rows of electrode plates and highly efficient membrane filter bags. In this case, the technology benefits from good synergism between the ESP and FF during bag cleaning resulting in very high performance levels, small sizes and operational flexibility.

4.4  Costs

As with most control technologies, the costs of PM controls involve both capital and operating costs. A cost-effectiveness indicator such as $/ton as is typically used for other technologies (e.g. NO_x and SO₂) is very difficult to address for generic PM control costs, as the range of PM reductions for different fuels and processes is wide that cost ranges become useless. An attempt to summarize costs in terms of capital and O&M components is presented below.

Capital

While it is customary to indicate capital costs on a $/kW basis for power generation applications, this is not relevant for non-power applications since no electricity is generated. However, one of the main parameters dictating the “sizing” and hence, the costs of a PM control device, is the quantity of flue gas it must handle. As a result, it is more appropriate to generalize capital costs on a “$/ACFM” basis. The following values represent typical costs for several of these technologies (these numbers reflect unit sizes ranging from utility-size units up to about 2,000,000 ACFM to smaller process down to about 10,000 ACFM))

•  Dry ESPs - $15 - $40/ACFM

•  Wet ESPs - $15 - $40/ACFM

•  Reverse Air Fabric Filter - $17 - $40/ACFM

•  Pulse Jet Fabric Filter - $12 - $40/ACFM

•  Venturi Scrubber - $5 - $20/ACFM

•  Cyclone - $1 - $5/ACFM

O&M

O&M costs are difficult to generalize for such a variety of technologies and applications, as they are affected by many parameters that include type of fuel, type of process, local ash disposal options, local cost of power, etc. O&M costs include fixed costs (FOM) and variable costs (VOM). The costs provided below are presented in $/year-ACFM and reflect costs for coal based fuels but should reasonably apply to other sources as well.

•  Dry ESPs - $0.25 - $0.65/yr-ACFM

•  Wet ESPs - $0.15- $0.50/yr-ACFM

•  Reverse Air Fabric Filter - $0.35 - $0.75/yr-ACFM

•  Pulse Jet Fabric Filter - $0.50 - $0.90/yr-ACFM

•  Venturi Scrubber - $0.25 - $0.65/yr-ACFM

•  Cyclone – Not applicable

•  Dry ESPs - $0.45 - $0.60/yr-ACFM

•  Wet ESPs - $0.25 - $0.50/yr-ACFM

•  Reverse Air Fabric Filter - $0.70 - $0.80/yr-ACFM

•  Pulse Jet Fabric Filter - $.90 - $1.1/yr-ACFM

•  Venturi Scrubber - $1.2 - $1.8/yr-ACFM

•  Cyclone – Not applicable

5  MULTI-POLLUTANT CONTROL TECHNOLOGIES

Emerging environmental issues and proposed federal legislation (President’s Clear Skies Initiative, Carper Bill, Jeffords’ Bill) as well as state legislation (examples include MA, NY, NC, NH, CT) have driven interest in multi-pollutant (as opposed to single pollutant) control technologies capable of addressing air pollutant emissions more comprehensively with greater flexibility and ultimately lower cost. Multi-pollutant control technologies integrate in-situ and/or post-combustion controls of at least two of the following pollutants: SO₂, NO_x, and Hg (and other hazardous air pollutants including cadmium, arsenic, and nickel), and CO₂. Multi-pollutant controls are intended primarily for large utility coal-fired boilers since the complexity of some of these processes as well as regulatory drivers often limit them to larger, utility boilers. Since coal-fired boilers represent the single largest source category for both NO_x (as well as SO₂ and Hg) and PM in the thirteen-state region, it is worth considering some of these technologies.

5.1  Proposed Multi-pollutant Emission Regulations from Utility Boilers

In 2002 and 2003 three “multi-pollutant” bills were introduced in the US Congress that call for coordinated reductions in NO_x, SO₂, and Hg from coal-fired power plants [26]. Some of the bills also include emission limits for CO₂. The three bills are briefly summarized here.

•  The Clean Power Act (CPA, Jeffords) would amend the CAA to require electric power generation sources greater than 15 MW. It is the most stringent of the three proposals. It will cap SO₂ emissions at 2.26 mm TPY in 2008 (0.28 mm TPY in the western region that includes WRAP states and MT, WY and CA; and 1.98 mm TPY in the eastern region). For NO_x, the cap of 1.51 mm TPY is to be met by 2008. The cap on Hg is at 5 TPY, also to be met by 2008. In addition, this bill sets a cap of 2.08 billion TPY for CO₂ to be met by 2008 (roughly 1990 levels). Except for Hg, national trading will be allowed to meet the caps.

•  The Clear Skies Act (CSA) has been proposed by the Bush administration. It is the least stringent of the three proposals. It would cap SO₂ emissions at 4.5 mm TPY in 2010 and at 3 mm TPY in 2018. The corresponding limits for NO_x are 3 mm TPY (in 2008) and 1.7 mm TPY in 2018. For Hg, the proposed national caps are at 26 TPY in 2010 and 15 TPY in 2018. There are no limits for CO₂. A national trading program similar to the existing trading program for SO₂ emissions under Title IV of the Clean Air Act will be the implementation mechanism to achieve these caps. All electric generation sources greater than 25 MW would fall under this program.

•  The Clean Air Planning Act (CAPA, Carper,Breaux, Baucus, and Chafe)) was intended as middle ground between the CPA and CSA. For SO₂, the caps are 4.5 mm TPY by 2008, 3.5 mm TPY by 2012, and 2.25 mm TPY by 2015. The caps for NO_x are 1.87 mm TPY by 2008 and 1.7 mm TPY by 2012. The Hg cap limits are 24 TPY by 2008, and a potential cap of 5-16 TPY by 2012 (this cap to be set by EPA and implies a control in the range of 79 to 93% from current Hg emission level). Cap and trade program will be the implementation mechanism for all four pollutants, except trading for Hg will be limited. In a “hybrid” approach, limited trading for Hg would be allowed (each plant will be required to reduce its Hg emissions at site by at least 50% in 2008 and by 75% in 2012). For CO₂, CAPA proposes to stabilize CO₂ emissions at 2005 levels (approximately 2.6 billion TPY) by 2008, and then stabilize to 2001 levels (approximately 2.4 billion TPY) by 2012.

All three bills recognize and incorporate the WRAP SO₂ trading program by setting separate caps on SO₂ emissions in the West. The CPA and CAPA allow nationwide trading of NO_x, while the CSA divides the country into two zones for NO_x trading. The western zone includes the ND, SD, NE, OK, KS, western TX, the eleven states west of the Rockies, AK, HI and the U.S. territories. The largest differences among the three bills are in the Hg emissions reduction requirements. The first-phase Hg emissions caps under CSA and CAPA are about the same, but compliance would come two years earlier under CAPA. CPA has the most stringent Hg reduction requirement: a cap of 5 TPY or about 90% control. The CSA would allow nationwide Hg trading, while the CAPA would allow partial trading. There is no trading under CPA.

Both CSA and the first phase of CAPA have modest Hg emission reduction targets; these would make it possible in some cases to achieve reduction of Hg as a “co-benefit” of other control technologies, for example, from the combination of an SCR and wet scrubber. If one of these bills were enacted, there might be some additional incentive to install an SCR and/or FGD on plants for which there might not be justification on the basis of a single pollutant.

In terms of Hg co-benefits, the West is at a disadvantage as compared to the East. In the latter region, more utilities burn bituminous coals that are high in chlorine (which tends to increase the amount of oxidized Hg in flue gas) and in sulfur. Wet scrubbers are effective for the removal of oxidized Hg, but ineffective for removal of the elemental Hg that is the predominant form of Hg in many western power plants. If all coal-fired power plants must reduce Hg emissions by upwards of 70%, the West will have a more difficult job than the East, owing to differences in coal composition. The bills that allow Hg trading (CSA and CAPA) would allow western power plants to deploy Hg control technology at plants were the highest emissions reductions are likely to be achieved.

If the CPA is enacted or if none of the three bills are enacted this year, it is likely that EPA will continue with the MACT process for Hg control, which does not allow trading and which will probably impose a Hg emission reduction target in the range of 70% to 90% (or an emission limit in the range of 0.2 to 0.6 pound of Hg per trillion Btu input). In this case, coal-fired power plants will have to look at application of activated carbon injection, the most mature technology for Hg control currently, or one of the multi-pollutant processes under development. Activated carbon injection may require adding additional particulate control equipment (such as a polishing baghouse with high cloth to air ratio), which will lower PM as well as the emissions of other hazardous pollutants including arsenic, chromium, lead, manganese, and nickel) as a consequence.

5.2  Multi-pollutant Control Technologies

A multi-pollutant control technology may be one integrated process or a combination of synergistic processes. In addition to in-situ and post-combustion control processes, options such as advanced power generation technologies, power plant rehabilitation-upgrading-repowering, fuel switching or blending and power plant optimization are sometimes included in the multi-pollutant control category. Emerging and commercial processes for multi-pollutant control for coal-fired boilers are summarized in Table 25, which is largely taken from Reference 4, with more recent information from the DOE-EPRI-U.S. EPA -A&WMA Combined Power Plant Air Pollutant Control Symposium in Washington, D.C., May 19-22, 2003.

Approximately half of the options listed in Table 25 are in commercial and early commercial stages. However, nearly all the options in commercial stage are proven SO₂ control technologies, which also remove Hg, advanced power generation options and power plant upgrading-fuel switching options. Nearly all in-situ and post-combustion controls (SO₂-NO_x or SO₂-NO_x-Hg) are either in demonstration or pilot-scale. Some technologies (e.g., SNO_X, SNRB, Advacate and CZD) have been tested either in pilot or demonstration scale in the early phase of the U.S. Department of Energy’s Clean Coal Technology (CCT) program, but have not been adopted by the industry. Some of these technologies may become more cost-effective if additional controls are required. Most of the environmental control processes increase the auxiliary power requirements of the plant (some up to 5%, but mostly in the range of 1 to 2%), increasing proportionally the CO₂ emissions.

Emerging post-combustion, multi-pollutant control technologies are being developed by a number of companies. The Electro-Catalytic Oxidation (ECO) system is a four-stage pollution control process that integrates established technologies to remove SO₂, NO_x, Hg and PM2.5. The system also produces a valuable fertilizer byproduct. The AIRborne process removes SO₂ and NO_x from plant emissions while turning the leftover material into a high-quality granular fertilizer. EnviroScrub is a dry scrubbing system that results in control of SO₂, NO_x, and possibly mercury and results in a byproduct that can be sold into the fertilizer, chemical, and/or explosives industry. None of these technologies controls emissions of CO₂.

Capital costs of options controlling two pollutants (either SO₂-NO_x or SO₂-Hg) are projected to be in the 50-315 $/kW range, but there is significant uncertainty associated with these estimates because of their early stage of development. Also, lack of information, especially associated with O&M costs, makes it difficult to compare their cost-effectiveness. Further monitoring and updating of cost-related information is needed. For reference, the costs of the combined commercial technologies, FGD and SCR are above 200-250 $/kW.

Advanced power generation technologies such as circulating fluidized bed (CFB), pressurized fluidized bed (PFBC) and integrated gasification combined cycle (IGCC) are potentially attractive options because they are revenue-generating options, while reducing significantly SO₂ and NO_x, and to a lesser extent CO₂. These options are available mainly for new power plants. Also, supercritical pulverized coal boiler provides an attractive alternative to subcritical pulverized coal boiler for nearly the same investment and results in an additional 4-12% reduction of all emissions. While this may not seem to be a significant percentage, their cost-effectiveness is attractive; also, the amount of CO₂ reduction (in tons or tons per year) is significant.

Of particular interest are options such as power plant optimization, fuel blending or switching and power plant upgrading. These options may play an important role in a flexible compliance regulatory framework and may result in significant savings for the utility industry compared to the implementation of control technology options. Optimization involves only operating changes, and while it results in only minor emission reductions, its costs are very low and therefore it is an attractive option and should be pursued in all power plants. Fuel blending or switching, and power plant upgrading provide significant opportunities for emission control, but their site-specific nature makes it difficult to generalize regarding their emission reduction potential and cost-effectiveness. A more site-specific assessment is recommended to assess the potential for these options in a typical utility system.

Issues	SO2/NOx/Mercury Control	High costs, especially operating costs due to high activated coke costs	Not widely demonstrated at full scale, ash salability, ESP/FF performance, impact of mercury speciation	Demonstration required	Demonstration required; costs estimated to be 30-50% lower than FGD-SCR	Depends on Hg speciation in flue gas.
Applicability		New and retrofit	Retrofit and new units with ESP an/or FF	New and Retrofit	New and retrofit.	Plants with SCR and Wet scrubber technologies
Emissions Reductions		SO2: 90-98%; NOx: 60-80%; Hg: 90-99%	Hg: 50-90%	SO2: 95-98%; NOx: 90%; Hg: 70+%	SO2: 99+%; NOx: 99%; Hg: 60-70%	SO2: 95%; NOx: 90-95%; Hg: 0-80%
Status		C	P/C	D	D	C
Technology		Activated Coke	Activated carbon with particulate controls	Electro Catalytic Oxidation	EnviroScrub	Wet FGD and SCR

6  SUMMARY AND RECOMMENDATIONS

6.1  NO_x and PM Sources

The main objectives of this project were to identify and briefly describe the available (or emerging) technologies for control of NO_x and PM emissions that could be applied to sources in the western United States. The starting point for this work was an analysis of large (greater than 100 TPY) sources from the WRAP 1996 Emission Inventory (Version 3). Sources were limited to those from the thirteen-state region: AZ, CA, CO, ID, MT, ND, NM, NV, OR, SD, UT, WA, and WY.

The source profile from the thirteen-state region was compared with that from the nine-state GCVTR: AZ, CA, CO, ID, NM, NV, OR, UT, and WY. The GCVTR accounted for 75% of the NO_x emissions and 83% of the PM emissions within the thirteen-state region. Generally, the distribution of sources was the same in the GCVTR as compared to the thirteen-state region. Thus, conclusions based on the thirteen-state region should therefore be valid for the GCVTR while achieving broader applicability to WRAP members.

The cut-off of 100 TPY captures 84% of the NO_x emissions in the 1996 WRAP database for the thirteen-state region. For ICE’s (reciprocating engines and turbines) the 100 TPY cut-off only captures about 56% of the emissions, though this category is the second largest category and responsible for 10% of stationary source emissions. Thus, NO_x control programs for sources in this category will require careful consideration of population attributes (e.g., a large number of small sources).

The largest source category for NO_x by far in the thirteen-state region is coal-fired boilers (68%); the top five categories (coal-fired boilers, internal combustion engines, cement kilns, turbines and oil and natural gas boilers) account for 90% of the NO_x emissions. The states with the largest stationary source NO_x emissions according to the 1996 WRAP database were AZ, CA, ND, NM, UT, and WY.

According to the WRAP 1996 (Version 3) stationary source emissions database, about 4% of the NO_x sources greater than 100 TPY had at least one type of control. Coal-fired boilers had the highest level of control (15%), followed by petrochemical processes (13%). The level of control for coal-fired boilers seemed low, even for 1996. Therefore, the 1996 WRAP database was compared with the data available for utility boilers in the 1996 CEMS and E-GRID databases. The EIA-767 database was also searched for NO_x control technologies. This comparison only looked at coal-fired utility boilers and not all coal-fired boilers. However, only 3% of the WRAP NO_x emissions from coal-fired boilers in the thirteen-state region were from non-utility boilers. WRAP data augmented by these other databases suggested that 44% of the utility boilers had at least one type of NO_x control in 1996, mostly low-NO_x burners.

The NO_x emission rate from external combustion boilers that is achievable with combustion modification depends on the fuel type. For coal-fired boilers, lower NO_x emission rates are obtained when firing subbituminous coal as compared to bituminous coal. Considering the amount of subbituminous coal in the West, there is a fairly even split between bituminous and subbituminous coals as fuels for utility boilers. This may have shifted since 1996, however.

The cut-off of 100 TPY captures 60% of the PM emissions in the 1996 WRAP database for the thirteen-state region. The largest source category in the thirteen-state region is coal-fired boilers (40%); the top eight categories account for 92% of the PM emissions (greater than 100 TPY): coal-fired boilers, mineral processing, petrochemical, wood boilers, fugitive, primary metal production, pulp and paper, and smelting operations. The state with the largest PM emissions is WY, followed by AZ, ID, and NM.

In the 1996 WRAP database, 72% of coal-fired boilers, the largest category of emissions, had PM controls. Overall, though, only 38% of units had PM controls.