Episodic Air Pollution
Control Measures:
.alysis of Potential
)tions for Industrial
/Sources
-^
January 2008
:
fi«*
&EPA
United States
Environmental Protection
Agency
-------
This page deliberately left blank.
-------
Episodic Air Pollution
Control Measures:
Analysis of Potential Options
for Industrial Sources
Report
January 2008
Prepared for:
U.S. Environmental Protection Agency
Office of Policy, Economics and Innovations
Sector Strategies Division
Work Assignment Manager: Barry Elman
elman.barry@epa.gov
Prepared by:
ICF International
9300 Lee Highway
Fairfax, VA 22031
(703) 934-3000
Project Manager: Bansari Saha
bsaha@icfi.com
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Table of Contents
List of Exhibits ii
List of Acronyms iii
Acknowledgements iv
Executive Summary ES-1
1. Introduction 1
I. Background on the Issue 1
II. Purpose of this Report 1
III. Selection of Sectors: Research Scope and Bounds 2
IV. Methodology, Data Sources, and Caveats 3
V. Report Outline 4
2. Sector Options and Analyses 5
I. Pulp & Paper 5
a. Overview of the Pulp & Paper Manufacturing Process and its Emissions 5
b. Location 7
c. Potential Episodic Control Measures 9
d. Opportunity Assessment of Potential Episodic Control Options 17
II. Iron & Steel 18
a. Overview of the Iron & Steel Manufacturing Process and its Emissions 18
b. Location 20
c. Potential Episodic Control Measures 22
d. Opportunity Assessment of Potential Episodic Control Options 27
III. Cement 28
a. Overview of the Cement Manufacturing Process and its Emissions 28
b. Location 30
c. Potential Episodic Control Measures 32
d. Opportunity Assessment of Potential Episodic Control Options 39
IV. Implications for Other Sectors 40
a. Implications for Industrial Boilers 40
b. Implications for Other Industrial Sources 42
c. Implications for Electric Utilities 43
3. Summary of Findings & Next Steps 44
I. Findings 44
II. Next Steps 46
Appendix: Examples of Reactions by Industrial Sectors
To the California Outages of 2001 and 2005 48
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
List of Exhibits
Exhibit 1: Point Source Emissions from Selected Manufacturing Sectors as a Percentage of All
Manufacturing Sectors
Exhibit 2: Basic Pulp & Papermaking Process
Exhibit 3: Primary Fuel Inputs as Fraction of Total Energy Supply for the Pulp & Paper Industry in 2002
(fuel use only)
Exhibit 4: Process Specific Emissions in 2002
Exhibit 5: Location of Pulp and Paper Mills in Relation to 8-Hour Ozone Nonattainment and Maintenance
Areas
Exhibit 6: Location of Pulp and Paper Mills in Relation to Counties Violating the 24-Hour PM2.5 NAAQS
Exhibit 7: Opportunity Assessment - Pulp & Paper
Exhibit 8: Basic Iron & Steel Production Process
Exhibit 9: Primary Fuel Inputs as Fraction of Total Energy Supply for the Iron & Steel Industry in 2002
(fuel use only)
Exhibit 10: Location of Iron and Steel Mills in Relation to 8-Hour Ozone Nonattainment and Maintenance
Areas
Exhibit 11: Location of Iron and Steel Mills in Relation to Counties Violating the 24-Hour PM2.5 NAAQS
Exhibit 12: Opportunity Assessment - Iron & Steel
Exhibit 13: Basic Cement Production Process
Exhibit 14: Primary Fuel Inputs as Fraction of Total Energy Supply for the Cement Industry in 2002 (fuel
use only)
Exhibit 15: Location of Cement Plants in Relation to 8-Hour Ozone Nonattainment and Maintenance
Areas
Exhibit 16: Location of Cement Plants in Relation to Counties Violating the 24-Hour PM2.5 NAAQS
Exhibit 17: Opportunity Assessment - Cement
Exhibit 18: Industrial Boiler Inventory - Boiler Capacity (Number of Boiler Units in Parenthesis)
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
List of Acronyms
AF&PA American Forest & Paper Association
AIRS Aerometric Information Retrieval System (EPA)
AISI American Iron and Steel Institute
AQM Air Quality Management (subcommittee of EPA's Clean Air Act Advisory Council)
EOF Basic oxygen furnace
Btu British thermal units
CAA Clean Air Act
CAAAC Clean Air Act Advisory Council (EPA)
CAP Criteria air pollutant
CKD Cement kiln dust
CO Carbon monoxide
DOE US Department of Energy
EAF Electric arc furnace
EPA US Environmental Protection Agency
ESP Electrostatic precipitator
FGD Flue gas desulfurization
CIS Geographic Information System
Ib Pound
LPG Liquified petroleum gas
MMBtu Million Btu
NAAQS National Ambient Air Quality Standard
NAICS North American Industrial Classification System
NEI National Emissions Inventory (EPA)
NESCAUM Northeast States for Coordinated Air Use Management
NESHAP National Emissions Standard for Hazardous Air Pollutant
NH3 Ammonia
NOx Nitrogen oxides
O&M Operation and maintenance
OAQPS Office of Air Quality Planning and Standards (EPA)
OPEI Office of Policy, Economics, and Innovation (EPA)
PCA Portland Cement Association
PM Particulate matter
PM2.5 Particulate matter that is 2.5 micrometers or smaller in size
PRB Powder River Basin (coal)
SCAQMD South Coast Air Quality Management District
SIC Standard Industrial Classification
SIP State Implementation Plan
SNCR Selective Non-Catalytic Reduction
SO2 Sulfur dioxide
SSD Sector Strategies Division (EPA)
TDF Tire derived fuel
TRI Toxics Release Inventory (EPA)
VOC Volatile organic compound
in
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Acknowledgements
We would like to thank the following individuals for providing technical information on the
industry sectors, pointing us to other experts and sources of information, and reviewing sections
of this report: Timothy Hunt of the American Forest & Paper Association; Arun Someshwar and
John Pinkerton of the National Council for Air and Stream Improvement; Lawrence
Kavanagh and Bruce Steiner of the American Iron & Steel Institute; Paul Sheehan of the
Sponseller Group, Inc. (Chair of the Environmental Technology Operating Committee of the
Association for Iron and Steel Technology); and Tom Carter of the Portland Cement
Association. Full responsibility for the contents of this report remains, however, with ICF
International and EPA.
IV
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Executive Summary
The Air Quality Management (AQM) Subcommittee of EPA's Clean Air Act Advisory Committee
(CAAAC) recently recommended expanding the use of episodic control measures to help attain
and maintain National Ambient Air Quality Standards (NAAQS) in areas where all reasonable
continuous (i.e., year-round) and seasonal control measures have already been required.1
Episodic control measures may be particularly useful in areas that are at risk of exceeding the
8-hour ozone NAAQS and/or the 24-hour fine particle (PM2.5) NAAQS only on a limited number
of days each year.
This study seeks to help EPA, states, communities, and other stakeholders assess the potential
use of episodic control measures by industrial sources. For purposes of this analysis, episodic
control measures are defined as measures that could be implemented intermittently, on 24-
hours advance notice, and for a total of up to 10 continuous or non-continuous days per year, to
help attain or maintain the short-term NAAQS for ozone or PM2.5.
Three industry sectors were selected as case studies for this analysis: pulp and paper; iron and
steel; and cement. We researched specific technological options for achieving additional
emission reductions on an episodic basis within the three selected sectors and assessed the
implications of these options for emission sources in other industry sectors. The options fell into
the following categories:
» Fuel switching, i.e., replacing current fuel sources with alternate fuels that result in lower
emissions of one or more priority pollutants;
» Low cost retrofits and enhanced use of existing control equipment, i.e., installing low
capital cost control devices, or using existing control devices more aggressively (or in
alternate ways);
» Scheduling changes, e.g., implementing short-term curtailments or targeted outages;
» Dispatch changes, e.g., shifting load from one boiler to another which has a more effective
control device or uses a less polluting fuel;
» Combustion optimization (or re-optimization), e.g., using sensors, controls or clean-
burning combustion modifications to reduce emissions;
» Sector-specific opportunities, e.g., increasing use of cogeneration units at pulp & paper
facilities.
We identified and assessed these options through literature reviews and web research on
manufacturing processes and control methods. The main information source for this study,
however, has been discussions with sector experts, including industry analysts at ICF and EPA,
members of trade associations, independent consultants, corporate environmental managers
and plant-level operators.
This study finds there are a number of potential episodic control measures available for each of
the three sectors analyzed.
» All three sectors possess some fuel switching capabilities and may be able to shift to lower
emitting fuels to reduce NOx, SO2, and/or PM on an intermittent basis for short periods of
time. The manner and extent to which short-term fuel switching could be employed,
1 See Recommendations to the Clean Air Act Advisory Committee: AQM Subcommittee Phase II Report,
June 2007, http://www.epa.gov/air/caaac/aqm/phase2finalrept2007.pdf.
ES-1
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
however, varies from sector to sector. Since facilities typically select their fuels so as to
minimize production costs and meet other operational objectives, switching to more
expensive fuels on a continuous basis would be untenable for many facilities. However,
because fuel switching under an episodic control program would only be invoked on a
limited number of days each year, the total incremental costs may be within acceptable
levels for many facilities.
Installation of low capital cost retrofits, such as selective non-catalytic reduction (SNCR)
and gas reburn systems (for NOx control), and lime injection systems (for SO2 control)
appears to hold promise in the context of an episodic control program. Although the high
operation and maintenance (O&M) costs associated with these retrofits (e.g., due to the
significant consumption of reagent or natural gas) may make them cost prohibitive for use
on a continuous basis at many facilities, they may be suitable for use at some facilities as an
episodic control measure.
Opportunities may exist to modify and thereby enhance the use of existing pollution
control equipment on a temporary basis to control for additional pollutants (i.e., in addition
to the main pollutant(s) these controls are intended for). For example, many industrial
boilers and furnaces have venturi scrubbers. These scrubbers are generally installed as PM
control devices but may also remove some SO2 from the exhaust stream of the combustion
unit. Some industry analysts and regulatory experts believe there may be an opportunity to
enhance the SO2 removal efficiency of venturi scrubbers in some situations by adding alkali
reagent. This may be technically feasible and economically viable where a unit already has
such a device and is able to obtain significant incremental SO2 reductions on a short-term
basis with the addition of the reagent.
Opportunities may also exist to run existing pollution control devices more aggressively
for short intervals to achieve incremental emission reductions. Some industry analysts
thought that in some cases it might be possible to temporarily increase the control efficiency
of a variety of NOx, SO2, and PM control systems on days when high air pollution
concentrations occur and where, for cost or other reasons, existing permit conditions do not
require these control devices to be utilized at their maximum performance levels.
While scheduling changes involving temporary shutdowns and curtailments that could
disrupt core production processes were generally deemed inadvisable, there may be
opportunities to curtail certain operations or rearrange production schedules on a short-term
basis to move high emitting processes to times when they are less likely to exacerbate air
quality problems. For example, because steel mini mill operation is generally batch in
nature, some mini mills may be able to reschedule melting operations from daytime to
nighttime hours to reduce VOC and NOx emissions, as well as reduce electrical demand,
during critical daylight hours when ozone formation is likely to occur. There may also be
opportunities to defer certain ancillary activities at cement plants, such as quarrying activities
and finishing mill operations, to reduce emissions as well as electrical demand on a short-
term basis.
Dispatch changes involving boilers at large pulp and paper mills with excess boiler capacity
appear to be a feasible episodic control strategy in some instances.
ES-2
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
» Additionally, there may also be other sector-specific opportunities for achieving additional
emission reductions on an episodic basis, such as increasing the use of cogeneration units
in the pulp and paper sector, which require further exploration.
» Episodic control measures related to combustion reoptimization techniques appeared less
viable for the selected industry sectors. Most facilities in these sectors use sophisticated
computer controlled equipment on highly inter-connected and synchronized production
processes. Industry experts were reluctant to entertain the idea of altering the already
optimized processes, as they might make the processes less efficient and/or more polluting.
While this study has focused primarily on facilities in the pulp and paper, iron and steel, and
cement manufacturing sectors, our research and information gathering indicates that a wide
variety of episodic control measures may be available for use by sources in other industrial
sectors as well.
» Several episodic control measures that are applicable to industrial boilers in the pulp and
paper or the iron and steel sectors, may also be broadly applicable to boilers in other
industrial sectors. For example, some industrial boilers in other industries may be able to
install low capital cost retrofit technologies (such as SNCR or gas reburn for NOx control, or
lime injection for SO2 control), to reduce boiler emissions on a short-term basis. Other
boilers may be able to temporarily reduce emissions by switching to a lower emitting fuel for
limited periods of time. In addition, some boilers that have already been retrofitted with
control devices may have the ability to increase the control efficiencies of those devices by
modifying their operation or running them more aggressively for short time intervals.
» As with industrial boilers, a number of the episodic control measures applicable to other
units and processes in the three selected industries may be applicable to similar units or
processes in other industries as well. Examples include fuel switching at furnaces and
process heaters, installation of low capital cost retrofits, enhanced use of existing control
equipment, and certain scheduling changes. Some sectors may also be able to provide
additional, sector-specific opportunities for achieving emission reductions on an episodic
basis.
« Several of the episodic control measures examined in relation to the three selected industry
sectors involve the rescheduling of electricity-intensive processes to times when they are
less likely to exacerbate air quality problems. Other industrial sectors may have similar
opportunities to temporarily halt, curtail or defer electricity-intensive processes as part of an
episodic control program. These measures could be especially helpful on high electricity
demand days when electric utilities may rely on very high emitting generating units in order
to meet peak demand for electricity. Reducing peak electricity demand, or shifting that
demand from daytime to nighttime hours (when lower-emitting generating capacity is
available and the resultant emissions may have less ozone or PM2.5 forming impact), could
provide significant air quality benefits. Electric utilities may also have the ability to employ a
variety of other episodic control measures to reduce NOx, SO2 and/or PM emissions on
days with high air pollution concentrations.
This report is intended to provide a preliminary assessment of the wide array of episodic control
measures that may be available to industrial sources. The potential measures identified in this
report should be studied further as EPA, states, communities, and other stakeholders consider
ES-3
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
the potential role that episodic control measures could play in helping to meet ozone and PM2.5
air quality goals. For example:
» To better assess the technical feasibility, emission reduction potential and cost of
prospective episodic control measures, it will be necessary to conduct detailed engineering
and cost analyses of specific measures. Because of the heterogeneity across facilities
within a sector, as well as across sectors, these analyses will need to take into account site-
specific and process-level emissions, as well as other source-specific considerations at
individual facilities.
« Because the ambient impact of episodic control measures is likely to be spatially
differentiated, location-specific air quality modeling will be necessary to determine the actual
impact that short-term emission reductions from any particular industrial source or sector
would have on local and/or regional air quality problems.
» In order to evaluate the potential utility of episodic control measures it will also be necessary
to review state-of-the-art practices pertaining to the prediction of ozone and PM episodes.
The effectiveness of episodic control measures would in part depend on the degree of prior
notice that can be provided to plant operators, which is in turn a function of the ability of
regulatory agencies and/or other institutions to predict air pollution episodes.
Another key step in the further exploration of episodic control measures is to quantify the cost-
effectiveness of these measures on a dollar-per-ton of pollutant removed basis. Our
discussions with industry analysts indicated this is a major unknown because episodic
measures for industrial sources have traditionally been overlooked in the air pollution control
literature. It must be emphasized that comparisons to cost-per-ton estimates based on
continuous control programs would provide an erroneous picture, as the acceptable cost-
effectiveness range for an episodic control program is likely to be substantially higher than the
corresponding range for a continuous program. Thus, a better understanding of the range of
cost-effectiveness values that could be appropriate given the benefits of an episodic control
program needs to be carefully developed.
A number of critical policy issues will also need to be addressed before episodic control
measures can be systematically implemented, including whether these measures should be
mandated by regulation, negotiated as part of the permitting process, or pursued by individual
facilities on a purely voluntary basis.
ES-4
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
1. Introduction
I. Background on the Issue
The Air Quality Management (AQM) Subcommittee of EPA's Clean Air Act Advisory Committee
(CAAAC) recently recommended expanding the use of episodic control measures to help attain
and maintain National Ambient Air Quality Standards (NAAQS) in areas where all reasonable
continuous (i.e., year-round) and seasonal control measures have already been required.2 A
number of communities have already developed programs to reduce emissions by individuals
and businesses (e.g., from vehicle use, road construction, open burning and other activities), on
specific days when high ozone or particulate matter (PM) concentrations are expected. To date,
however, few efforts have been made to apply episodic control measures to industrial sources.3
Industrial source episodic control measures—measures that can be implemented intermittently
for short periods of time—could provide an expanded set of control options for states and
communities. These options may be particularly useful in areas that are at risk of exceeding the
8-hour ozone NAAQS and/or the 24-hour fine particle (PM2.5) NAAQS only on a limited number
of days each year. A variety of measures that could not be implemented on a continuous or
seasonal basis could potentially prove suitable and acceptable for episodic use. By reducing
peak concentrations on the highest pollution days, episodic control measures can provide
considerable health and environmental benefits.4
II. Purpose of this Report
The Sector Strategies Division (SSD) within EPA's Office of Policy, Economics, and Innovation
(OPEI) undertook this study to help EPA, states, communities, and other stakeholders assess
the potential use of episodic control measures by industrial sources. Three industry sectors
were selected as case studies for analysis: pulp and paper; iron and steel; and cement.
This study builds on a previous analysis conducted by ICF International in 2006 to identify and
qualify a subset of industry sectors as candidates for further episodic control research. Our goal
in this current study is to identify, through research and discussions with industry experts
nationwide, potential episodic control measures for the three selected sectors that may be
technically feasible, practical, and cost-effective, and to assess the implications of these
approaches for other industry sectors.
For purposes of this analysis, episodic control measures are defined as measures that could be
implemented intermittently, on 24-hours notice, and for a total of up to 10 continuous or non-
continuous days per year, to help attain and maintain NAAQS for ozone or PM2.5.
2 See Recommendations to the Clean Air Act Advisory Committee: AQM Subcommittee Phase II Report,
June 2007, http://www.epa.gov/air/caaac/aqm/phase2finalrept2007.pdf.
3 Memorandum from Art von Lehe to Barry Elman, EPA Sector Strategies Division, Re: Episodic Control
Measures, August 11, 2006
4 The AQM Subcommittee recognized that legal restrictions currently limit EPA's ability to provide State
Implementation Plan (SIP) credit for stationary source episodic control measures. The Subcommittee did,
however, recommend expanding the use of such measures as "backup insurance mechanisms" (outside
the scope of an approved SIP) for areas working to attain or maintain the short-term ambient air quality
standards.
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
III. Selection of Sectors: Research Scope and Bounds
In 2006, ICF International conducted an analysis for SSD to identify industry sectors that
warrant further study with regard to their ability to use episodic control measures to help states
and local communities meet their ozone and PM2.5 attainment goals.5 We used three metrics
as screens for that analysis: plant location vis-a-vis nonattainment designations, emissions of
relevant criteria air pollutants, and energy consumption.
« Plant Location. Using a Geographic Information System (GIS) framework, we analyzed
sectors based on their geographic distribution and correspondence with current short-term
ozone and PM2.5 nonattainment designations. Information on nonattainment status came
from the EPA Green Book.6 The resulting visual representation allowed us to determine
sectors that could potentially provide significant episodic emission reductions in
nonattainment areas.
» Criteria Air Pollutant (CAP) Emissions. We analyzed air emissions profiles for all 28
industry sectors included in the 2001 Aerometric Information Retrieval System (AIRS) and
estimated their relative contributions to total CAP emissions (i.e., VOC, NOx, SO2, and
PM2.5) that contribute to ozone and PM2.5 air quality problems.
» Energy Consumption. Building on SSD's on-going analysis of strategies to promote
environmentally-preferable energy outcomes in selected manufacturing sectors,7 we
analyzed the energy intensities and fuel mix for those selected sectors to determine whether
there may be short-term changes in fuel use and combustion equipment that may offer
significant potential for episodic emission reductions.
Based on these screening assessments, we identified a number of sectors as good candidates
for further study of potential episodic control measures, including the pulp and paper, iron and
steel, and cement sectors. Section 2 of this report provides updated location maps for these
three sectors, while Exhibit 1, below, provides updated emissions data.8
5 "Selecting Sectors for Analysis of Potential Episodic Control Measures", ICF technical memorandum to
Barry Elman, EPA Sector Strategies Division. September 29, 2006.
6 See http://www.epa.gov/oar/oaqps/greenbk/index.html.
7 See March 2007 final report, Energy Trends in Selected Manufacturing Sectors: Opportunities and
Challenges for Environmentally Preferable Energy Outcomes, prepared by ICF for SSD,
http://www.epa.gov/ispd/energy/index.html.
8 While the three sectors selected for case studies all have manufacturing facilities located in ozone
and/or PM2.5 nonattainment areas, and they all emit significant quantities of criteria pollutants and
operate energy intensive processes, further analysis (including assessment of site-specific emissions,
stack parameters, air chemistry, atmospheric transport and other factors) would be necessary in order to
fully characterize the air quality impacts of any individual facility or industry sector.
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
Exhibit 1: Point Source Emissions from Selected Manufacturing
Sectors as a Percentage of All Manufacturing Sectors
Sector
Pulp & Paper*
Iron & Steel**
Cement***
All Manufacturing Sectors****
VOC
8%
1%
1%
100%
NOX
14%
4%
15%
100%
S02
15%
3%
8%
100%
PM2.5
12%
7%
6%
100%
* Pulp and Paper is defined as NAICS codes 32211, 322121 and 322122, or SIC codes 261 and 262
** Iron & Steel is defined based on a list of facilities developed by EPA's Sector Strategies Division
*** Cement is defined based on a list of facilities developed by EPA's Sector Strategies Division
**** All Manufacturing Sectors is defined as NAICS codes 31-33, or SIC codes 20-39
Source: U.S. EPA's Final v2 2002 National Emission Inventory (NEI) for Criteria Air Pollutants from Point
Sources
IV. Methodology, Data Sources, and Caveats
Building upon our 2006 analysis, we researched specific technological options for achieving
intermittent, short-term emission reductions across the target sectors. These options fell into the
following categories:
« Fuel switching, i.e., replacing current fuel sources with alternate fuels that result in lower
emissions of one or more priority pollutants;
» Low cost retrofits and enhanced use of existing control equipment, i.e., installing low
capital cost control devices, or using existing control devices more aggressively (or in
alternate ways);
» Scheduling changes, e.g., implementing short-term curtailments or targeted outages;
» Dispatch changes, e.g., shifting load from one boiler to another which has a more effective
control device or uses a less polluting fuel;
» Combustion optimization (or re-optimization), e.g., using sensors, controls or clean-
burning combustion modifications to reduce emissions;
» Sector-specific opportunities, e.g., increasing use of cogeneration units at pulp & paper
facilities.
We identified and assessed these options through literature reviews and web research on
manufacturing processes and control methods. The main information source for this study,
however, has been discussions with sector experts, including industry analysts at ICF and EPA,
members of trade associations, independent consultants, corporate environmental managers
and plant-level operators. Within the constraints of time and resources, we conducted a limited
set of interviews, seeking information and expert opinions to vet the technical feasibility,
practicality and cost-effectiveness of these potential options. As a result, some information is
anecdotal, and there are technical points on which knowledgeable opinions vary.
Clearly, consideration of episodic control measures for industrial sources is in its formative
stage. This report is intended to provide a preliminary basis for assessing the wide array of
options that may be possible. Based on this analysis, however, we cannot say definitely which
measures will or will not be viable, or what benefits they will provide. The potential measures
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
identified in this report should be studied further as EPA, states, communities and other
stakeholders consider options for achieving ozone and PM2.5 attainment goals.
V. Report Outline
The major sections of this report are organized as follows:
« Section 2, Sector Options and Analyses, explores potential episodic control measures for
each of the three sectors selected as case studies for this analysis, and considers the
applicability of these measures to other industrial sources.
o The first three subsections present the results of our research on the three selected
sectors: 2.1, pulp and paper; 2.11, iron & steel; and 2.Ill, cement. We begin each sector's
discussion by presenting an overview of its manufacturing processes and related
emissions, as well as by providing maps that show the location of the sector's facilities in
relation to nonattainment areas for ozone and PM2.5. We then provide our detailed
analysis of potential episodic control measures, followed by summaries of our main
findings in text boxes.
o Section 2.1V highlights the implications of our analysis for other sectors, focusing on
potential measures for industrial boilers, as well as for electric utilities and other
industrial sources.
« Section 3, Summary of Findings & Next Steps, summarizes our findings across the three
selected sectors, as well as the implications for other sectors, and presents consolidated
next steps.
» The Appendix describes the effects that the 2001 and 2005 rolling blackouts in California
had on the performance of several industry sectors, to inform discussion on how these or
other sectors might react to unscheduled outages in the context of an episodic control
program.
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
2. Sector Options and Analyses
I. Pulp and Paper
a. Overview of the Pulp & Paper Manufacturing Process and its Emissions
Pulp and paper manufacturing can be broken down into five significant steps:9
» Wood preparation involves removing the bark from logs and then breaking down the
debarked logs into wood chips of uniform size. The uniformity in the chip size maximizes the
quality and efficiency of the subsequent pulping process.
« Pulping then converts the fibrous wood material into a slurry of fibers. Pulping involves
cooking the wood fibers, washing, and screening to remove unwanted materials. Three
different types of pulping technologies are used, depending on the desired properties of the
final product and the corresponding amount of lignin that needs to be removed from the
paper.10 Chemical pulping removes the most lignin, semichemical pulping removes some
lignin, while mechanical pulping does not remove any lignin.
« Chemical recovery, which is part of the pulping process, enables the recovery and reuse of
chemicals used in the chemical and semichemical pulping process. Steam and electricity
generated during the recovery process also help offset the large energy requirements of
pulp and papermaking.
» Bleaching is a chemical process used to whiten or brighten the pulp before it is used in
papermaking. Different bleaching techniques are used depending on the specific pulping
process and the residual lignin left in the pulp.
» Papermaking is the final stage consisting of four steps:
o Preparing a homogeneous pulp slurry (stock);
o Dewatering the slurry;
o Pressing and drying to manufacture the paper; and
o Finishing the manufactured paper. Depending on the final product desired,
finishing can involve one or more of the following processes - rewinding the
paper onto a reel, trimming, coating, printing, saturation, and box-making.
Exhibit 2 presents a simple flow diagram of the basic pulp and paper manufacturing process,
applicable for all three types of pulping processes - chemical, semichemical, and mechanical
pulping.
9 The following overview of the pulp and paper manufacturing process and emissions is taken from the
DOE/ITP report on the pulp and paper industry; (Energy and Environmental Profile of the U.S. Pulp and
Paper Industry, US DOE, Office of Energy Efficiency and Renewable Energy, Industrial Technologies
Program, December 2005). This overview describes the process at an integrated pulp and paper mill.
The industry also has a large number of non-integrated mills that purchase pulp to make paper and
others that obtain and process recycled fibers for papermaking.
10 Wood lignin holds the wood fiber together and adds strength and stiffness to trees, but also results in
weaker paper that yellows with age.
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
Exhibit 2: Basic Pulp & Papermaking Process
COOKING
WOODYARD AND CHIPPING
Source: Smook, 1992 (from EPA Sector's Notebook, November, 2002)
Exhibit 3 below shows the relative shares of the different fuel types used in the pulp and paper
industry in 2002. The "Other" fuel category, which primarily includes biomass, is the dominant
fuel at about 54 percent of the total. Biomass fuel includes spent liquor or black liquor11
(approximately 70 percent of the "Other" category) and wood residues and byproducts
(approximately 27 percent of the remainder). The next biggest fuel source is natural gas at 21
percent, followed by coal and electricity at 10 percent and 9 percent, respectively.
Exhibit 3: Primary Fuel Inputs as Fraction of Total Energy Supply for the
Pulp & Paper Industry in 2002 (fuel use only)
Coal
10%
Other
54%
Net Electricity
9%
Natural Gas
21%
Fuel Oil
5%
Source: Energy Trends in Selected Manufacturing Sectors, EPA Sector Strategies Division, March
2007, based on DOE/EIA Manufacturing Energy Consumption Survey, 2002 Data Tables. Totals may
not add to 100 percent due to rounding.
Net electricity provides a rough approximation for purchased power and includes purchased
electricity, transfers in, and generation from noncombustible renewables, and excludes electricity
transferred out or sold.
Black liquor is a byproduct generated in the chemical pulping process. Boilers fired with black liquor
have an efficiency of about 70 percent, compared to 80-85 percent for coal- or natural gas-fired boilers of
same size. (Source: Characterization of the U.S. Industrial/Commercial Boiler Population, Oak Ridge
National Laboratory, May 2005.)
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
Air Emissions
The biggest emission sources in a pulp and paper mill are the power boilers used to generate
steam and electricity for sustaining the pulp and papermaking process. These power boilers
consume all of the coal used by this sector and produce significant amounts of SO2 (about 83
percent of the total for this industry) and NOx (about 77 percent of the total), and some VOC.
The second biggest emission source is the kraft recovery furnace. While recovery furnaces
generate significantly lower SO2 and NOx emissions than the power boilers, their VOC
emissions are similar to those from the power boilers. Power boiler PM emissions are generally
well controlled through venturi scrubbers, baghouses or ESPs, while ESPs are by far the
dominant PM control device on kraft recovery furnaces. Exhibit 4 presents process-specific
emissions from the pulp and paper industry.
Exhibit 4: Process Specific Emissions in 2002
Pulp and Paper Emissions
•annnnn
ocnnnn
81"
900000
k.
a> 1 snnnn
c 1 nnnnn
njOOOO
o
n, lln-« n_
ifl . = «{L =
Source
nS02
• NOx
nvoc
D Total HAP
• PM-10PRI
Source: SO2 and NOx Emissions from U.S. Pulp and Paper Mills, 1980-2005. National Council for Air
and Stream Improvement (NCASI) analysis (based on Draft 2002 NEI data).
b. Location
There were 425 pulp and paper mills, including 148 integrated mills, operating in the U.S. in
2005.12 The location of mills and facilities appears to be largely driven by the location of wood
sources, rather than energy sources. For example, pulp mills that use virgin woods tend to be
in regions where trees are harvested, including the Southeast, Northwest, Northeast, and North
Central regions. Pulp mills that rely on recycled fiber are located near sources of waste paper.
Nonintegrated paper mills often are located near pulping operations, since this is their primary
input.13 Exhibit 5 below presents the location of pulp and paper mills in relation to
nonattainment and maintenance areas for the current 8-hour ozone NAAQS, as of October
12
Pulp and Paper Mill Emissions of SO2, NOx, and Particulate Matter in 2005, Special Report No. 06-07,
National Council for Air and Stream Improvement (NCASI), December 2006.
13 Analysis of Energy Usage in Selected Industrial and Commercial Sectors, US EPA, Sector Strategies
Division, May 24, 2006.
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
2QQ7 14 15 Exhibit 6 presents the location of pulp and paper mills in relation to counties violating
the current 24-hour PM2.5 NAAQS, based on 2004-06 monitoring data.16 While these exhibits
indicate that a number of pulp and paper mills are located in or near areas that currently violate
the short-term standards for ozone or PM2.5, it should be noted that many of these areas are
projected to attain the standards in the near term while other areas are at risk of violating the
standards in the future. It should also be emphasized that source-specific factors, including
process-level emissions, stack parameters, topography, meteorology, and air chemistry, would
need to be taken into account when assessing the air quality impact of any individual facility or
sector on local or regional air quality problems.
Exhibit 5: Location of Pulp and Paper Mills in Relation to
8-Hour Ozone Nonattainment and Maintenance Areas
LEGEND
• Pulp & Paper Mills
] 8-Hour Ozone Nonattainment Areas
8-Hour Ozone Maintenance Areas
Note Partial nonsmginment counties (i.e., Ihose wilh port
Sources: Pulp and paper mill locations obtained for facilities listed with SIC codes 261 and 262, or
NAICS codes 32211, 322121 and 322122, in 2002 (v2) National Emissions Inventory (NEI) CAP
Point Source Dataset. Nonattainment and maintenance areas for the current 8-hour ozone NAAQS
obtained from EPA's Green Book: Nonattainment Status for Each County by Year, October 2007.
14 The locations of pulp and paper mills were obtained for facilities listed with SIC codes 261 and 262, or
NAICS codes 32211, 322121 and 322122, in 2002 (v2) National Emissions Inventory (NEI) CAP Point
Source Dataset. Data on nonattainment and maintenance areas for the current 8-hour ozone NAAQS
were obtained from EPA's Green Book: Nonattainment Status for Each County by Year, October 2007,
available at http://www.epa.gov/oar/oaqps/greenbk/map8hrnm.html and
http://www.epa.gov/oar/oaqps/greenbk/anay.html.
15 EPA proposed tightening the 8-hour ozone NAAQS on July 11, 2007. See:
http://www.epa.gov/air/ozonepollution/pdfs/20070711_proposal_fr.pdf.
16 The locations of pulp and paper mills were obtained for facilities listed with SIC codes 261 and 262, or
NAICS codes 32211, 322121 and 322122, in 2002 (v2) National Emissions Inventory (NEI) CAP Point
Source Dataset. Data on counties violating the current 24-hour PM2.5 NAAQS were obtained from EPA's
Air Quality System (AQS) as of July 11, 2007, available at http://www.epa.gov/airtrends/values.html.
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
Exhibit 6: Location of Pulp and Paper Mills in Relation to
Counties Violating the 24-Hour PM2.5 NAAQS
LEGEND
• Pulp & Paper Mills
^1 Counties Violating the 24-hour PM2 5 NAAQS
Note: A number of .idcltlion.il counties, fai which complete maniloring
data were not yet available, may also be designated as violating the 24-
hour PM2.5 NAAQS once the monitoring clala are complete.
Sources: Pulp and paper mill locations obtained for facilities listed with SIC codes 261 and 262, or
NAICS codes 32211, 322121 and 322122, in 2002 (v2) National Emissions Inventory (NEI) CAP
Point Source Dataset. Counties violating the current 24-hour PM2.5 NAAQS (based on 2004-06
monitoring data) obtained from EPA's Air Quality System (AQS) as of July 11, 2007.
c. Potential Episodic Control Measures
We focused our analysis on five categories of potential episodic control measures for the pulp
and paper sector:
• Fuel switching;
• Low cost retrofits and enhanced use of existing control equipment;
• Dispatch changes;
• Scheduling changes; and
• Cogeneration.
/'. Fuel switching
Several short-term fuel switching options appear to have potential as episodic control measures
for the pulp and paper industry:
The pulp and paper industry uses a large number of power boilers and coal is a major fuel
source for many of these boilers. Coal accounts for a disproportionate share of the SO2 and
NOx emissions from power boilers, and there appears to be some potential for switching from
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
coal to other lower emitting fuels, such as biomass or natural gas, for short time periods in many
power boilers. 17
Most power boilers in this industry are "combination boilers" capable of burning multiple fuel
types. Combination boilers can shift between fuel types on short notice and are generally
designed to run up to 100 percent on any of the fuels they are designed to handle. According to
industry analysts, many combination boilers routinely burn coal and biomass, with the
percentage breakdown between the two fuel sources varying over time and across facilities.
Typically, facilities that burn biomass have at least a one week onsite supply of biomass fuel.
Thus, facilities burning both coal and biomass in a combination boiler may have the flexibility to
increase their use of biomass, or even to use biomass as the only fuel source, for short periods,
if needed. Industry analysts felt that if biomass was used to displace coal there should be
significant SO2 and NOx emission reduction benefits. Further analysis is, however, needed to
confirm that facilities with combination boilers have the ability to shift on short (i.e., 24-hours)
notice to a higher share of biomass in their fuel mix (i.e., up to 100 percent biomass) for short
periods of time, since current permit limitations and/or operational constraints could restrict such
changes. Further analysis is also needed to determine the magnitude of the SO2 and NOx
emission reductions that short-term fuel switching from coal to biomass could provide, as well
as the impact on other pollutants, and to more fully assess the cost and operational impacts on
individual facilities.
Another potential short-term fuel switching option for power boilers is switching from coal to
natural gas. Switching to natural gas on a short-term basis appears to be technically viable for
many facilities since most boilers have natural gas firing capability. The key issue with regard to
this fuel switching option appears to be cost. Facilities may have to make limited capital
expenditures to prepare for using natural gas, and would have to establish operational
procedures, train staff and maintain equipment on an ongoing basis in order to ensure
"operational readiness" to switch over to natural gas on short notice. Moreover, facilities would
need to secure a reliable natural gas supply, and this would likely entail a significant price
premium (over the already high cost of natural gas) in order to ensure sufficient access to
natural gas on short notice. Nevertheless, some industry analysts thought it was conceivable
that the cost of shifting to natural gas on a limited number of days with high air pollution
concentrations could be less than the cost of installing further add-on controls. This would need
to be confirmed through further analysis, as would the magnitude of the SO2 and NOx
reductions, and the operational impacts on individual facilities, that would result from
implementing this strategy.
In addition, industry analysts also noted that some pulp and paper mills in the Midwest use high
sulfur coal and may be good candidates for switching to lower sulfur coal on days with high air
pollution concentrations, such as coal from the Powder River Basin (PRB). Although such a
switch might entail additional fuel and/or transportation costs, the proximity of these mills to PRB
in Wyoming and lower rail tariffs resulting from railway deregulation may make this switch a
viable and cost-effective episodic control strategy for some mills.18 Pulp and paper mills that
employ this strategy would need to stockpile and manage a secondary coal supply, which would
17 The combustion of coal and biomass both result in higher PM emissions than the combustion of natural
gas and together they account for most of the PM emitted from power boilers in the pulp and paper
sector.
18 Industry analysts pointed out there may also be costs associated with equipment modifications needed
to burn subbituminous versus bituminous coal and there may be technical feasibility constraints to the use
of PRB coal in cyclone boilers.
10
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
further increase operating costs and could potentially raise other logistical or operational issues.
Moreover, unlike electric utilities, pulp and paper facilities are not primary purchasers of coal
and may be restricted by existing long-term coal contracts.19 Therefore, they may be
constrained in their ability to obtain (or replenish) low sulfur coal supplies. Nevertheless,
switching from high- to low-sulfur coal on short (i.e., 24 hours) notice warrants further
exploration as a potentially viable and cost-effective episodic control strategy for some
Midwestern mills that currently burn high sulfur coal.
Another fuel switching option that might work for smaller mills is switching from residual oil to
natural gas. Most small mills have package boilers designed to burn fuel oil and/or natural gas.
According to industry analysts, many of these boilers currently burn residual oil but can switch to
natural gas on short notice. This switch can provide significant SO2 and NOx emission reduction
benefits. Switching from residual oil to natural gas may be a particularly attractive option for
small mills located in or near urbanized areas experiencing air quality problems. There are
several concerns, however, with burning more natural gas at small facilities; these concerns
relate to accessibility, availability, and cost.
» Some small mills may not have access to a natural gas pipeline and would therefore be
unable to switch over to this fuel on days with high air pollution concentrations.
» Because natural gas is in high demand for heating purposes during the winter, as well as on
high electrical demand days during the summer, particularly in the Northeast where a
number of small mills are located, issues related to availability and price would need to be
assessed to determine the true viability of this option for small mills.20
For small mills with package boilers that burn residual oil, an alternative strategy could be to
switch to LPG or propane instead of natural gas. Our discussions with industry analysts
indicated that some mills might be able to store a limited supply of LPG onsite for use on days
when high air pollution concentrations occur. Alternatively, with a 24-hour lead time, mills might
be able to arrange to have LPG tanker(s) brought onsite to supply enough LPG to replace
residual oil for a short period of time (e.g., 8-12 hours). Our discussions indicated that this
might be a viable and cost-effective strategy for a number of small mills given the nature of their
operations, even though modest capital expenditures may be required (e.g., to install new
burners) before a boiler would be equipped to burn LPG.21 Further research is needed on the
costs and operational impacts of implementing this strategy. In addition, permitting restrictions
may need to be addressed in order to allow mills to store and/or use LPG as an episodic control
measure.
19 Primary purchasers of coal are those facilities/industries that are the main buyers of coal, and may
therefore have some influence on the type of coal being shipped as well as delivery schedules.
20 Mills might need to pay a price premium to secure a guaranteed, as-needed supply.
21 Switching from residual oil to distillate oil or diesel fuel may also be feasible with the installation of new
burners and a separate fuel storage tank.
11
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Summary: Fuel Switching
There appear to be several potential fuel switching options for achieving significant short-term SO2
and NOx emission reductions from power boilers as part of an episodic control strategy for the pulp
and paper industry.
For bigger mills, the most promising short-term fuel switching opportunities would occur at
combination boilers. Many of these boilers use large quantities of coal and may have the capability to
temporarily use more biomass or natural gas in its place to achieve significant emission reductions.
There might also be opportunities to switch from high- to low-sulfur coal for short intervals at some
Midwestern mills.
For small mills, short-term shifts from residual oil to natural gas or LPG/propane in package boilers
seem to hold promise. This strategy might be particularly useful where small mills are located in or
near urbanized areas with air quality problems.
/'/'. Low-cost retrofits and enhanced use of existing control equipment
Another set of episodic control measures we considered entails the use of add-on control
equipment. Although most pollution control retrofits require significant capital expenditures for
installation, there are some that have relatively low installation costs but higher operation and
maintenance (O&M) costs. Such low-capital, high O&M retrofits may be cost-effective to
operate for short time periods during an episodic air pollution event. Similarly, where control
technologies are already deployed, it may be possible to run them more aggressively for short
intervals, or in other ways that would not be feasible or cost-effective on a continuous basis.
We identified four control technologies that seemed to have potential as episodic control
measures for the pulp and paper industry:
« Selective non-catalytic reduction (SNCR)
« Gas reburn
« Venturi scrubbers
« Lime injection
SNCR for NOx Control
A number of industry analysts suggested that the use of SNCR systems may be a viable
episodic control measure for reducing NOx emissions from facilities in the pulp and paper
industry, particularly for base-loaded boilers (i.e., boilers that run continuously, or close to it).
SNCR systems can be used with base-loaded boilers firing any fuel (coal, wood, gas, etc.).
Some analysts have stated, however, that if the boiler load fluctuates significantly, there may be
technical problems with installing SNCR systems on such boilers. According to these analysts,
SNCR is not a proven technology on industrial boilers with fluctuating loads. However, one
SNCR vendor we spoke with has stated that SNCR has been demonstrated on many industrial
boilers with fluctuating load and that the cycling boilers in the pulp and paper industry do not
present special technological challenges.22 Thus, the use of SNCR systems as an episodic
control measure holds promise for base-loaded boilers in the pulp and paper industry. However,
22 This vendor said that it would guarantee the performance of SNCR on cycling boilers at pulp and paper
mills, although site-specific modeling would be required before a guarantee could be issued with regard to
any specific installation.
12
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
the extent to which these systems can be used as an episodic control measure for cycling
boilers would need to be further analyzed to reconcile the differing viewpoints.23
In those instances where the use of SNCR systems is technically feasible, the limiting factor is
generally the cost of implementing these systems. According to industry analysts and
technology vendors, the cost of purchasing and installing an SNCR system may be on the order
of $1 million or more for a large industrial boiler. However, for large boilers, most of the total
annualized costs associated with an SNCR system is for the reagent used (urea or aqueous
ammonia). Therefore, the cost burden associated with installing and operating an SNCR
system could be reduced significantly under an episodic control program. For example, in the
case of a large boiler it might be possible to reduce the total cost burden of an SNCR system by
50 percent or more if the system were installed and run at an intermediate (e.g., 30 percent)
control level on a continuous basis, and then ramped up to full capacity (e.g., 50 percent control
efficiency) on days with high air pollution concentrations. The cost burden could be further
reduced if the SNCR system were kept on standby and used only on days with high air pollution
concentrations.
A major concern expressed by industry analysts was that once an SNCR system is installed and
available for use as part of an episodic control program, regulators and/or community activists
may demand that it be run at full capacity on a continuous basis. Further analysis of the
technical feasibility, emission reduction potential and cost of this option with respect to different
boiler types, fuels and operating conditions is required.
Gas Reburn for NOx Control
Another NOx control technology that might be viable as an episodic control measure for this
sector is natural gas reburn. According to industry analysts, stoker wood-fired boilers in the
pulp and paper industry may be ideal candidates for additional NOx control using intermittent
gas reburn. Since most facilities may have access to natural gas (for use as a boiler starter
fuel, for example), the incremental costs of using gas reburn may be limited to the costs of
installing additional burners and using natural gas. Because the strategy would only be used
intermittently and for short periods, however, mills would require relatively limited quantities of
natural gas. Thus, the total incremental costs may be within acceptable levels for many large
mills.24
Venturi scrubbers for SO2 Control
About 25 percent of the combination boilers in the pulp and paper industry that burn both coal
and wood have venturi scrubbers. These scrubbers are installed for PM control but may also
have the capability to further reduce SO2 emissions by as much as 80 percent when alkali
reagent is used. Some mills already use alkali reagent in their venturi scrubbers to reduce SO2
emissions, however, the vast majority do not. The main reason why venturi scrubbers have not
been widely utilized as an SO2 abatement strategy is because they are generally not considered
to be efficient or cost-effective for that purpose - according to one industry analyst they may
waste about half the reagent used.25 The use of alkali reagent to achieve additional SO2
23 Another issue needing further evaluation according to some industry analysts is the potential for
ammonium sulfate formation in the case of boilers with high sulfur fuels.
24 Although natural gas reburn has fallen out of favor as a continuous control strategy in the current
scenario of high gas prices, its use as an episodic strategy may still be a worthwhile avenue of research.
25 Facilities that need to scrub SO2 on a continuous basis would generally install packed tower scrubbers
or other SO2 control technologies that are more efficient in their use of alkali reagent, and are capable of
achieving SO2 reductions of 90 percent or more.
13
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
emission reductions may, however, be technically feasible and cost-effective as an episodic
control measure for those facilities that have already installed venturi scrubbers, insofar as they
could achieve a significant incremental reduction in SO2 emissions on a short-term basis for a
modest cost with the addition of alkali reagent.26
Lime injection for SO2 Control
Mills that use baghouses or electrostatic precipitators (ESP) for PM control may have the option
to inject dry lime powder into the duct that carries exhaust gases from a power boiler to the PM
control device, to get additional SO2 removal. The costs for implementing this control strategy
could include the cost of installing storage, transfer and injection systems for the lime, and the
cost of the lime reagent itself. For kraft mills with existing baghouses or ESPs on their boilers,
this strategy may be particularly attractive and cost-effective because lime is already available
at the mill (although the particle size distribution that is needed for lime injection may not be
easily available). Moreover, in the case of baghouses, caking of the lime on the baghouse walls
would provide enhanced control efficiency for SO2, with over 70 percent control efficiencies
possible according to one expert. For mills that use ESP for PM control, the absence of caking
on the walls would imply that their control efficiencies would be lower, thereby increasing the
cost-per-ton of SO2 removed. Even with ESPs, however, lime injection may still be able to
achieve significant short-term SO2 reductions at modest cost.
Summary: Low Cost Retrofits and Enhanced Use of Existing Control Equipment
Most industry analysts agree that it is technically feasible to install SNCR systems on base-loaded
boilers in the pulp and paper industry. Because of the relatively low capital costs and the limited use
of reagent, installing and using SNCR systems as part of an episodic NOx control program may be a
cost-effective strategy for these boilers. There is some dispute, however, as to the technical feasibility
of installing SNCR systems on pulp and paper industry boilers with vacillating load. Further analysis
of the technical feasibility and cost of this option with respect to different types of boilers and operating
conditions is required.
Another NOx control technology that may warrant further consideration as an episodic control
measure is natural gas reburn. Although the high price of natural gas makes this technology less
appealing for use on a continuous basis, it may be cost-effective for use as an intermittent, short-term
control strategy.
The addition of alkali reagent to venturi scrubbers to achieve further SO2 reductions appears to be a
promising episodic control measure for mills that have already installed these scrubbers for PM control
purposes. These mills would need to incur the additional cost (O&M) of using reagent only on a
limited number of days when high air pollution concentrations occur (as well as the cost of installing an
alkali addition system for those mills that do not presently have such a system). Its effectiveness,
however, would vary depending on the fuel mix used in the boilers and other site-specific factors.
For those mills that use baghouses or ESPs for PM control, lime injection may be a viable and cost-
effective episodic SO2 control measure.
26 The SO2 reduction resulting from the addition of alkali reagent would also depend on the coal/wood fuel
mix in the boilers. If the boiler burns mostly coal and very little wood (say 90-10 ratio), then adding
reagent can provide up to 80 percent SO2 reduction.
14
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
/'/'/'. Dispatch changes
According to industry analysts, a typical large, integrated mill may have several power boilers,
e.g., two large combination boilers and two smaller package boilers designed to burn fuel oil
and/or natural gas. One or more of these boilers would likely be base-loaded and the others
cycled as needed. In some cases, an integrated mill may also have an idle package boiler that
could be brought online to burn natural gas with as little as one hour of lead time. Thus, at any
given moment, a mill may have some excess boiler capacity, and this capacity could occur at a
boiler that burns cleaner fuel or has a more efficient control device than a boiler that is currently
being utilized. Most boilers within a facility are inter-connected so that load can be shifted from
one to another. Therefore, from a theoretical standpoint, mills may have some ability to dispatch
different boilers at different times to attain episodic emission reduction goals. Further research
is needed, however, to determine the magnitude of the emission reductions that could result
from short-term dispatch changes, and to more fully assess the cost and operational impacts on
individual facilities.
Another possible dispatching strategy for pulp and paper mills involves generating more steam
at recovery furnaces, on a short-term basis, to reduce the load on power boilers. Most of the
steam that is required at an integrated mill is provided by burning black liquor in the recovery
furnace. Although the recovery furnace is typically operated at a high utilization rate, there may
be some ability to further increase its capacity utilization on a temporary basis. For example, a
recovery furnace that is operating at 85 percent of its maximum capacity could be ramped up to
90 or 95 percent for a short interval in order to reduce the demand for steam from boilers,
thereby reducing the need to run the boilers to generate steam. Based on our discussions, it
appears that recovery furnaces can be readily cycled up or down, and most mills have a
sufficient stockpile of black liquor to allow for a one or two day increase in the furnace's
utilization rate.27 Because black liquor has lower fuel heat content than coal, more black liquor
would be needed to generate the same amount of energy. Nevertheless, due to the lower SO2
and NOx emissions generally resulting from black liquor combustion compared with most fossil
fuels, it may be possible to achieve a net reduction in SO2 and NOx emissions. Further
research is needed to determine the magnitude of the emission reductions that could be
achieved by increasing steam generation at the recovery furnace on a short-term basis, and to
more fully assess the cost and operational impacts on individual facilities.
Summary: Dispatch Changes
Changing boiler dispatch as an episodic control strategy may work for some integrated pulp and paper
mills to the extent that they have excess capacity at low emitting boilers on days with high air pollution
concentrations. In addition, integrated mills may be able to temporarily increase steam production at
their recovery furnace in order to reduce their reliance on higher emitting power boilers to generate
steam.
27 In addition, many recovery furnaces have load-carrying oil or gas burners that could be used to produce
additional steam. Use of auxiliary oil would not increase SO2 emissions due to capture of the oil sulfur by
sodium fumes inside the furnace. Additional NOX from auxiliary gas or oil burning in the furnace would
still be less than the NOX produced from coal combustion on a Ib/lb steam produced basis. However,
there could be air permitting issues associated with increased auxiliary fuel burning.
15
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
iv. Scheduling changes
Short-term scheduling changes appear to be the least viable episodic strategy for pulp and
paper mills, as they almost always strive to operate at constant capacity in order to maintain
maximum production efficiency. Industry analysts indicated that short-term changes to mill
operations would likely reduce production efficiency and therefore be cost prohibitive.
Moreover, short-term disruptions in mill operations could increase emissions intensity
(emissions per unit of production), and potentially increase net emissions from a facility.
Summary: Scheduling Changes
Industry analysts agreed that scheduling changes would generally not be effective as an episodic
strategy for pulp and paper mills.
v. Cogeneration
The pulp and paper industry is one of the largest cogeneration applications among all industry
sectors. Cogeneration is considered an environmentally preferable generating technology,
because the simultaneous production of thermal and electric energy from the same fuel source
is more efficient than generating just electricity. Onsite electricity generation also eliminates
energy losses associated with transmission and distribution of power over the electric grid.
According to industry analysts, a significant portion of the pulp and paper industry's
cogeneration is considered emission neutral, because it uses primarily the steam generated
from the pulping process. More than 65 percent of the industry's electricity needs are met
through cogeneration processes.28
Most analysts we spoke with agreed that the pulp and paper industry may have the ability to
further increase its electricity generation through its cogeneration systems on a short-term basis
(i.e., provide a "surge capacity"), but this would need to be studied further. Moreover, any
strategy that involves increasing electricity generation from the pulp and paper industry for short
time intervals would need to evaluate the tradeoffs associated with displacing electricity
generated by power plants. While it is possible that increased electricity generation at pulp and
paper mills could displace generation from high emitting electrical utility units on high electricity
demand days, further research is needed to determine the extent to which net emission
reductions would result from this strategy, and to more fully assess the cost and operational
impacts on individual mills.
Summary: Cogeneration
Pulp and paper mills already lead all industry sectors in cogeneration, but they may still have some
excess capacity to increase their electricity generation on days when high air pollution concentrations
occur. However, any strategy to increase the pulp and paper industry's electricity generation would
need to consider the tradeoffs associated with displacing electricity generated by power plants.
Moreover, concerns exist in how increased cogeneration would impact the pulp and paper industry's
relationship with power plants and electricity rates. More analysis is needed to determine whether this
can be a viable episodic strategy.
28 Analysis of Energy Usage in Selected Industrial and Commercial Sectors, US EPA, Sector Strategies
Division, May 24, 2006.
16
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
d. Opportunity Assessment of Potential Episodic Control Options
Exhibit 7 summarizes the potential viability of the five primary options for further reducing
emissions on days with high air pollution concentrations.
Exhibit 7: Opportunity Assessment - Pulp & Paper
Option
Fuel switching
Low cost retrofits and
enhanced use of existing
control equipment
Dispatch changes
Scheduling changes
Increased cogeneration
Viability as Potential Episodic Control Measure
Promising - Several short-term fuel switching options appear to have
potential as episodic control measures (for SO2 and NOx reductions)
Promising - SNCR for base-load boilers (for NOx reductions); venturi
scrubbers and lime injection (for SO2 reductions)
Needs further study - Gas reburn (for NOx reductions)
Possible - For large mills with multiple power boilers and/or a recovery
boiler operating below maximum capacity on days when high air pollution
concentrations occur (for SO2 and NOx reductions)
Unlikely
Needs further study
17
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
II. Iron & Steel
a. Overview of the Iron & Steel Manufacturing Process and its Emissions
As detailed in Exhibit 8, the iron and steel manufacturing process can be broken down into the
following four major steps: cokemaking, ironmaking, steelmaking, and forming and finishing.29
« Cokemaking uses coal to generate coke, coke oven gas, and by-product chemicals from
compounds released from the coal. Coke is produced by heating metallurgical coal (also
called "coking coal") in such a manner as to drive off the volatile substances commonly
present in coal, producing a carbon-rich coke.
« In the traditional method of ironmaking, iron-bearing materials are reduced to molten
metallic iron (called "pig iron") in a blast furnace, using carbon present in coke as a reducing
agent. The iron-bearing materials are pellets or sinter made from iron ore and other
materials at ore agglomeration facilities. The pellets or sinter are charged with coke in the
blast furnace, and hot blast air is blown through tuyeres in the lower part of the furnace. The
hot gas from coke combustion creates a reducing atmosphere in which the iron oxides are
converted and melted to molten pig iron. Metallic iron also can be produced without using
coke in a direct reduction process, generating the necessary reducing atmosphere directly
with coal or natural gas. Pig iron contains 3 to 4 percent carbon and direct-reduced iron
commonly contains 1 to 3 percent carbon.
» Steelmaking removes carbon and other impurities by controlled oxidation of the iron
charged into a steelmaking furnace. In addition to hot metal from a blast furnace, iron and
steel scrap or direct-reduced iron can be charged into a steelmaking furnace. Additional
elements, chromium or nickel for example, may be added to steel in the furnace or in the
ladle after the furnace to produce special alloy steels. The principal steelmaking processes
used in the US are:
o Basic oxygen furnace (BOF) - molten iron from the blast furnace is refined by
injecting high purity oxygen and adding fluxing materials. BOF offers superior
energy efficiency with high throughput. Steelmaking facilities that use this
technology are supplied with molten iron from a blast furnace, along with up to 30
percent steel scrap. These facilities, which may also include a coke plant and
finishing facilities, are collectively known as integrated mills.
o Electric arc furnace (EAF) - steel scrap is melted and refined using electric energy,
so no cokemaking or ironmaking operations are required.30 EAF permits charging
with high levels of scrap, which reduces or eliminates the need for access to hot
metal from the blast furnace. Steel facilities that use this technology are referred to
as mini mills.
29 The following discussion of the basic iron and steelmaking process is summarized from two sources
EPA Sector Notebook on Iron and Steel, 1995, and Gas Research Institute, Environmentally Driven
Threats and Opportunities for Natural Gas in the Iron and Steel Industry, (prepared by ICF/EEA),
November 1994.
30 Some EAF facilities use scrap substitutes (e.g., pig iron or direct reduced iron) as well as scrap.
18
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
In the Forming and Finishing stage, once the molten steel has attained the desired
chemical composition in the steelmaking furnace, it is tapped (poured) into a ladle.
Additional chemical adjustment may occur in the ladle with the addition of alloying material
or the removal of unwanted constituents from the molten steel. The next step is the casting
of the molten steel. There are two types of casting used today: continuous casting and ingot
casting. Continuous casting accounts for about 97 percent of total steel produced. In
continuous casting, molten steel is shaped directly from the ladle to form semi-finished steel
products known as blooms, billets, and slabs. These semi-finished steel products are rarely
used as final products. They usually serve as raw materials in the manufacturing of final
steel products. The semi-finished products undergo various processes including hot rolling,
cold rolling, forging, and drawing to form the finished products.
Exhibit 8: Basic Iron & Steel Production Process
EXHIBIT*
IRQH AND MANUFACTURING
PROCESS OVERVIEW
Source: EPA Sector Notebook
Exhibit 9 shows the relative share of the different fuel types used in the iron and steel sector in
2002. Coke and breeze generated in the manufacturing process was the dominant fuel source
at 36 percent. The industry also used significant amounts of natural gas (27 percent). The
"Other" fuel category (21 percent) was largely composed of byproduct fuels such as coke oven
19
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
gas and blast furnace gas. (All of the coke and breeze and most of the fuels classified as
"other" are coal-based in origin.)
Exhibit 9: Primary Fuel Inputs as Fraction of Total Energy Supply for the Iron &
Steel Industry in 2002 (fuel use only)
Coal
3%
Other*
21%
Net Electricity
13%
Natural Gas
27%
Coke & Breeze*
36%
Source: Energy Trends in Selected Manufacturing Sectors, EPA Sector Strategies Division, March
2007, based on DOE/EIA Manufacturing Energy Consumption Survey, 2002 Data Tables. Totals may
not add to 100 percent due to rounding.
Net electricity provides a rough approximation for purchased power and includes purchased
electricity, transfers in, and generation from noncombustible renewables, and excludes electricity
transferred out or sold.
*AII of the coke and breeze, and most of the fuel inputs classified as "other" (i.e., byproduct fuels such
as coke oven gas and blast furnace gas), are coal-based in origin.
Air Emissions
The cokemaking process is a large source for air emissions, including fine particles of coke
generated during the process and sulfur compounds emitted from the coke oven stacks.
However, because the expulsion of volatile substances from coal is the purpose of cokemaking,
fugitive VOCs are the most important source of air emissions. These emissions include various
complex hydrocarbons, such as benzene and polycyclic aromatics, as well as other VOCs.
Ironmaking is also significant from an air pollution perspective, as it generates significant
amounts of NOx, some SO2, and particulates.31 Air emissions are generated from the crushing
and handling of the ore and from fuel combustion in the furnace. Ore crushing is the largest
source of uncontrolled PM in the iron and steel industry. The emissions from crushing
operations, while considerable, are effectively controlled by conventional PM technologies.
Steelmaking generates some amounts of NOx as a by-product of CO combustion in the exhaust
of both furnaces. The EAF process also generates additional NOx at the arc. Emissions of
VOCs and other pollutants depend on the quality of material charged into the furnace. VOC
emissions are derived primarily from the scrap charge, which is greater in the EAF.
b. Location
There are currently 18 integrated steel mills operating in the U.S., mostly in the Great Lakes
region. In addition, there are mini mills operating in approximately 70 locations across the
country, with a large concentration in the Great Lakes region as well.32 Exhibit 10 below
presents the location of iron and steel mills in relation to nonattainment and maintenance areas
for the current 8-hour ozone NAAQS, as of October 2007.33 34 Exhibit 11 presents the location
31 Energy and Environmental Profile of the U.S. Iron and Steel Industry, US Department of Energy, Office
of Industrial Technologies, August 2000.
32 List of iron and steel facilities prepared by U.S. EPA Sector Strategies Division.
33 The locations of iron and steel mills were obtained from the American Iron and Steel Institute. Data on
nonattainment and maintenance areas for the current 8-hour ozone NAAQS were obtained from EPA's
Green Book: Nonattainment Status for Each County by Year, October 2007, available at
http://www.epa.gov/oar/oaqps/greenbk/map8hrnm.html and
http://www.epa.gov/oar/oaqps/greenbk/anay.html.
20
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
of iron and steel mills in relation to counties violating the current 24-hour PM2.5 NAAQS, based
on 2004-06 monitoring data.35 While these exhibits indicate that a number of iron and steel
mills are located in or near areas that currently violate the short-term standards for ozone or
PM2.5, it should be noted that many of these areas are projected to attain the standards in the
near term while other areas are at risk of violating the standards in the future. It should also be
emphasized that source-specific factors, including process-level emissions, stack parameters,
topography, meteorology and air chemistry, would need to be taken into account when
assessing the air quality impact of any individual facility or sector on local or regional air quality
problems.
Exhibit 10: Location of Iron and Steel Mills in Relation to
8-Hour Ozone Nonattainment and Maintenance Areas
LEGEND
Iron & Steel Mills
* Integrated Mills
• Mini Mills
__ 8-Hour Ozone Nonattainment Areas
I 1 8-Hour Ozone Maintenance Areas
Note: Partial nonattainmenl counties (i.e., those with part
of the county designated as nonallainmone and part as
attainment) are shown as nDnallainmenl on the map.
Sources: Iron and steel mill locations obtained from U.S. EPA Sector Strategies Division.
Nonattainment and maintenance areas for the current 8-hour ozone NAAQS obtained from EPA's
Green Book: Nonattainment Status for Each County by Year, October 2007.
34
EPA proposed tightening the 8-hour ozone NAAQS on July 11, 2007. See:
http://www.epa.gov/air/ozonepollution/pdfs/20070711_proposal_fr.pdf.
35 The locations of iron and steel mills were obtained from the American Iron and Steel Institute. Data on
counties violating the current 24-hour PM2.5 NAAQS were obtained from EPA's Air Quality System (AQS)
as of July 11, 2007, available at http://www.epa.gov/airtrends/values.html.
21
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
Exhibit 11: Location of Iron and Steel Mills in Relation to
Counties Violating the 24-Hour PM2.5 NAAQS
LEGEND
Iran & Steel Mills
* Integrated Mills
• Mini Mills
CZi Counties Violating the 24-hour PM2 5 NAAQS
Note: A number ol additional counlies, (or vrfiich complete monilonng
data were not yet available, may also be designated as violating the 24-
hour PM2.5 NAAOS otioe the monitoring data are complete
Sources: Iron and steel mill locations obtained from U.S. EPA Sector Strategies Division. Counties
violating the current 24-hour PM2.5 NAAQS (based on 2004-06 monitoring data) obtained from EPA's
Air Quality System (AQS) as of July 11, 2007.
c. Potential Episodic Control Measures
We focused our analysis on four categories of potential episodic control measures for the iron
and steel sector:
• Fuel switching;
• Low cost retrofits and enhanced use of existing control equipment;
• Scheduling changes; and
• Dispatch changes.
/'. Fuel switching
For integrated iron and steel mills that produce coke, little opportunity exists to switch from coal
to other fuels in the coke-making process. Since coke-making is one of the most polluting
processes in this industry, integrated mills may, however, reduce their reliance on coke by
burning pulverized coal as an auxiliary fuel in the blast furnace. This would decrease the
amount of coke they need to produce and yield a corresponding decrease in emissions from the
coke-making process. There are technical limits, however, on the extent to which pulverized
coal can be used to displace coke in a blast furnace. Moreover, once a mill has made the
considerable capital investment necessary to use pulverized coal, it would be expected to use
that fuel (which is less expensive than coke) to the maximum extent possible. Therefore, there
appears to be little opportunity to increase the use of pulverized coal in a blast furnace as part of
a strategy to reduce coke oven emissions on days with high air pollution concentrations. Some
22
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
blast furnaces may, however, be able to use natural gas as an auxiliary fuel on an episodic
basis. Using natural gas would likely result in NOx, SO2, and PM emission reductions. More
research is needed, however, to assess the economic and technical feasibility of this option.
Most boilers at integrated iron and steel mills are designed to burn multiple fuels and are able to
switch among these fuels on a few hours notice. Consequently, another short-term fuel
switching option that might hold some promise for reducing NOx, SO2, and VOC emissions at
integrated mills is to switch from coke oven gas to natural gas in boilers on days with high air
pollution concentrations. Most integrated mills have some capability to store unused coke oven
gas but must flare any unused gas that they are unable to store.36 Depending on the storage
capacity of their gas holders, and how much of that capacity is unutilized at the time, mills may
be able to increase the use of natural gas in their boilers for a limited period of time, while
saving the displaced coke oven gas for later use. As discussed below, mills may also be able to
slow the operation of their coke ovens as part of an episodic control strategy, which would
decrease the amount of coke oven gas that is available for use in the boilers. This shortfall
could then be made up by burning more natural gas. Further analysis is needed to determine
the extent to which integrated mills could shift from coke oven gas to natural gas in boilers, and
the degree to which such shifts would reduce NOx, SO2, and VOC emissions.37 Emission and
cost impacts would need to be assessed on a plant-by-plant basis.
Because mini mills do not have any significant fuel needs and depend primarily on grid-supplied
electricity, there do not seem to be any fuel switching options for mini mills.
Summary: Fuel Switching
Integrated mills may be able to temporarily slow down the coke-making process to decrease coke
oven emissions and decrease the supply of coke oven gas to the boilers. The boilers could
compensate by burning more natural gas which would result in reduced boiler emissions of NOx, SO2,
and VOC. There may also be opportunities to store a limited amount of coke oven gas in gas holders,
which would enable mills to temporarily increase their use of natural gas at boilers while saving the
displaced coke oven gas for later use.
Some integrated mills may also be able to use natural gas as an auxiliary fuel in the blast furnace to
achieve short-term reductions in NOx, SO2, and PM emissions. This could also allow for a temporary
decrease in coke production and a corresponding decrease in emissions from the coke-making
process.
These potential fuel switching options for integrated mills need to be studied at the plant-level, as their
infrastructural needs, operational impacts and subsequent costs may vary by plant.
No fuel switching options were identified for mini mills.
36 Indications are that most integrated mills may have gas holders to store limited quantities of unused
coke oven gas (e.g., one or two days supply).
37 Our initial discussions with industry analysts
benefit if mills were to switch from coke oven gas to natural gas.
37 Our initial discussions with industry analysts seemed to point towards a significant emission reduction
23
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
/'/'. Low-cost retrofits and enhanced use of existing control equipment
Another set of episodic control measures we considered entails the use of add-on control
equipment. Although most pollution control retrofits require significant capital expenditure for
installation, there are some controls that have relatively low installation costs but higher
operation and maintenance (O&M) costs. Such low-capital, high O&M retrofits may be cost-
effective to operate for short periods when high air pollution concentrations occur. Similarly,
where control technologies are already deployed, it may be possible to run them more
aggressively for short intervals, or in other ways that would not be feasible or cost-effective on a
continuous basis.
« We identified one control technology that seemed to have potential as an episodic control
measure for iron and steel mills: Selective Non-Catalytic Reduction (SNCR).
SNCR for NOx Control
Some analysts we spoke with felt that SNCR may hold promise as a short-term control measure
for reducing NOx emissions from integrated iron and steel mills. While SNCR technology has
been in use for many years in some industries, no SNCR system has yet been installed at an
iron or steel mill. The limiting factor has been the cost of implementing these systems.
According to industry analysts and technology vendors, the cost of purchasing and installing an
SNCR system may be on the order of $1 million or more for a large industrial boiler. However,
for large boilers, most of the total annualized cost associated with an SNCR system is for the
reagent used (urea or aqueous ammonia). Therefore, the cost burden associated with installing
and operating an SNCR system could be reduced significantly under an episodic control
program. For example, in the case of a large boiler it might be possible to reduce the total cost
burden of an SNCR system by 50 percent or more if the system were installed and run at an
intermediate (e.g., 30 percent) control level on a continuous basis, and then ramped up to full
capacity (e.g., 50 percent control efficiency) on days with high air pollution concentrations. The
cost burden could be further reduced if the SNCR system were kept on standby and used only
on days with high air pollution concentrations.
A major concern expressed by industry analysts was that once an SNCR system is installed and
available for use as part of an episodic control program, regulators and/or community activists
may demand that it be run at full capacity on a continuous basis. Further analysis of the costs
of this option and other site-specific considerations is required.
Summary: Low Cost Retrofits and Enhanced Use of Existing Control Equipment
Some analysts felt that an SNCR system may hold promise as an episodic control measure for reducing
NOx emissions from large boilers at integrated mills and warrants further research. Because of the
relatively low capital cost and the ability to limit reagent consumption, installing and using SNCRs as part
of an episodic control program may be a cost-effective strategy for this industry.
/'/'/'. Scheduling changes
Integrated mills
Industry analysts felt there might be opportunities to curtail cokemaking operations at integrated
facilities for short periods in order to reduce VOC, NOx, SO2, and PM emissions. Facilities can
generally lengthen the coking cycle (e.g., from 16 hours to up to 24 hours) without disrupting
other parts of the production process, where they have sufficient onsite coke inventories or the
24
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
ability to buy additional coke from merchant coke manufacturers. There is indication that some
facilities may already have permit provisions that require them to use longer coking cycles in
response to upset conditions. As discussed above, slowing down the coking cycle would also
reduce the production of coke oven gas, allowing facility operators to increase the use of natural
gas in boilers, which would yield additional NOx, SO2, and VOC emission reductions.
Integrated facilities might also be able to increase scrap usage in the basic oxygen furnace for
short periods of time and on short notice. Basic oxygen furnaces typically charge between 10
and 30 percent scrap, with the permissible range varying from product to product. Within this
range, the choice of hot metal versus scrap is primarily based on price. Therefore, there may
be limited opportunity at some mills to temporarily shift to a higher scrap utilization rate within
the permissible range for a product, or to switch to production of a product that permits a higher
scrap utilization rate.
Another possible option would be to slow down or delay batch operations in the blast furnace for
short intervals. Industry analysts emphasized, however, that operational stability is very
important for blast furnaces. For this and other reasons, they believed that episodic control
measures that target scheduling changes in coke oven operations and/or increased scrap
utilization rates in the basic oxygen furnace may hold more promise for integrated steel mills
than scheduling changes affecting the blast furnace.
Options for rescheduling production activities at integrated mills need to be assessed with
regard to their potential costs and emission reduction benefits, and evaluated further with plant
operators to understand their effectiveness and the tolerable levels of disruption for integrated
mills.
Mini mills
Because mini mill operation is generally batch in nature, there may be opportunities to
rearrange production schedules on a short-term basis to move high emitting processes to times
when they are less likely to exacerbate air quality problems. For example, rescheduling melting
operations from daytime to nighttime hours could significantly reduce emissions of VOC and
NOx during critical daylight hours when ozone formation occurs. Similarly, rescheduling melting
operations from one time period to another could potentially shift PM and NOx emissions to a
period when those emissions would be less likely to contribute to a PM2.5 violation.
Since mini mills consume extremely large amounts of grid-supplied electricity, rescheduling
melting operations could be particularly useful on high electricity demand days when electric
utilities may utilize very high-emitting generating units to meet peak load requirements. Shifting
mini mill electricity consumption from daytime to nighttime hours (when lower-emitting
generating capacity is available and the resultant emissions may have less ozone or PM2.5
forming impact), could therefore provide significant air quality benefits.
Note that during the rolling blackouts in California in 2001 and 2005 (see the Appendix), some
mini mills with interruptible power contracts switched to firm contracts with their power suppliers
because the power curtailments were deemed too costly. Others, however, rescheduled their
melting operations to off-peak hours to take advantage of the cheaper electricity rates during
those times and still allow for the interruptible contracts to continue.38
38 Mini mills pointed out that an unscheduled stoppage that occurs during the course of the melting
operation is extremely dangerous to workers and could cause equipment destruction.
25
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
According to industry analysts, many mini mills in other parts of the country also have
interruptible power contracts so that electricity can be temporarily curtailed or halted when there
is a surge in electricity demand. There is also evidence that some mills in other parts of the
country run at night to take advantage of cheaper off-peak electricity rates.
Further analysis is needed to determine the extent to which mini mills may have additional
opportunities to shift production to off-peak electricity generation periods, and the magnitude of
the SO2 and NOx emission reductions that such short-term scheduling changes could provide.
Further analysis is also needed to more fully assess the cost and operational impacts of these
changes on individual facilities.
Summary: Scheduling Changes
Integrated steel mills might be able to implement longer coking cycles for short time periods if they have
adequate onsite coke inventories or access to merchant coke. These mills may also be able to increase
the use of scrap in their basic oxygen furnaces. Scheduling changes in blast furnaces, while technically
feasible, appear to be more disruptive for these facilities.
Due to the batch nature of mini mill operations, many mini mills already use interruptible power contracts,
or schedule their melting activities during off-peak hours, and there may be opportunities to make
additional scheduling changes to shift melting and other high emitting processes to times when they
would be less likely to exacerbate air quality problems. The deferral of melting operations from daytime
to nighttime hours could be particularly useful on high electricity demand days, since it may enable
utilities to avoid the use of very high-emitting generating units that would otherwise be needed to meet
peak load requirements.
iv. Dispatch changes
According to industry analysts, most integrated steel mills have at least two boilers with some
flexibility to shift load from one boiler to another on just a few hours notice. Analysts expressed
their doubts on whether such a change in boiler dispatch would be worthwhile, however, since
all boilers typically have the capability to use all the available fuel types, and fuel switching
within the same boiler may generally be a more effective option.
Summary: Dispatch Changes
Because integrated mills have access to multiple boilers, changing boiler dispatch may be theoretically
possible, but may not be practically effective. Most boilers can use all available fuel types, so switching
fuel within the same boiler may be more effective than changing boiler dispatch.
26
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
d. Opportunity Assessment of Potential Episodic Control Options
Exhibit 12 summarizes the potential viability of the four primary options for further reducing
emissions on days with high air pollution concentrations.
Exhibit 12: Opportunity Assessment - Iron & Steel
Option
Fuel switching
Low cost retrofits and
enhanced use of existing
control equipment
Scheduling changes
Dispatch changes
Viability as Potential Episodic Control Measure
Promising - Several short-term fuel switching options appear to have
potential at integrated facilities (for NOx, SO2, PM, and VOC reductions),
but coke oven gas storage options, and the ability to link slowdowns in
the coke-making process with increased use of natural gas at boilers,
need to be evaluated
Unlikely- At mini mills
Promising - SNCR on large boilers (for NOx reductions)
Promising - For both integrated facilities and mini mills
Needs Further Study - While it is likely that some integrated facilities
could shift load among boilers on short notice, it is unclear if this would
be an effective control measure
27
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
III. Cement
a. Overview of the Cement Manufacturing Process and its Emissions
Cement manufacturing requires the "thermochemical processing of substantial amounts of
limestone, clay, and sand in large kilns" at very high and sustained temperatures (over 2,750°F)
to produce an intermediate product called clinker.39 Clinker is ground up with small quantities of
gypsum to create Portland cement, the predominant variety of cement manufactured in the U.S.
Portland cement is used as a binding agent in virtually all concrete.40 The process of Portland
cement manufacturing can be broken down into the following four steps:41
» Quarrying and crushing the rock;
» Grinding the carefully proportioned materials to high fineness;
» Subjecting the raw mix to thermochemical processing in a kiln; and
« Grinding the resulting clinker to a fine powder.
Cement is manufactured in four different types of kilns:42
» Long wet kilns;
» Long dry kilns;
» Dry kilns with preheaters; and
» Dry kilns with precalciners.
Processes that take place within each kiln type include drying and preheating, which comprises
evaporation of free water, dehydration of clay minerals, and increasing the temperature of the
raw materials to over 1,650°F; calcining, a process of decomposing the carbon compounds,
which occurs when the raw materials reach about 1,650°F; and heating the calcined raw
materials to about 2,750°F, where they are sintered to form clinker. The same raw materials are
used in the wet and dry kilns, however, the moisture content and processing techniques differ,
along with kiln designs.43
« Wet process kilns are generally longer to allow them to dry the wet raw material mix (also
known as slurry) as it is fed into the kiln.
« Long dry kilns are more energy efficient than wet process kilns but produce exit gases at
very high temperatures.
» Preheater kilns differ from the traditional long wet and long dry kilns because they have a
series of cyclones where the dry raw material is intimately mixed with the hot combustion
gases. These kilns use about half of the fuel that a wet kiln uses to produce clinker because
the preheater system efficiently recovers process heat.
39 Assessment of Control Technology Options for BART-Eligible Sources, prepared by NESCAUM, March
2005.
40 Energy Trends in Selected Manufacturing Sectors, prepared by ICF for EPA's Sector Strategies
Division, March 2007.
41 Assessment ofNOx Emissions Reduction Strategies for Cement Kilns - Ellis County- Final Report,
prepared for Air Quality Planning Section, Texas Commission on Environmental Quality, July 14, 2006.
42 Ibid.
43 The following description of the cement manufacturing process and emissions profile is taken from the
EPA Sector Notebook on Stone, Clay, Glass, and Concrete Products Industry (including cement),
September 1995.
28
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
• Precalciner kilns have a calciner vessel inserted between the preheater and the kiln. In a
precalciner kiln about 40 percent of the fuel is fired in the kiln and the rest is fired in the
calciner. Firing a large amount of the fuel in the calciner significantly reduces thermal NOx
emissions because the fuel is burned at less than 2,000°F and normally fuel consumption for
a precalciner kiln is 10 percent less than a preheater kiln.
The clinker produced in the kiln is cooled in a traveling grate using ambient air and the cooled
clinker is transferred to storage. The clinker is then mixed with four to six percent gypsum and
ground to produce a homogeneous cement powder, which is typically sent to a bulk storage
area and then shipped by truck or rail.
Exhibit 13 presents a simple flow diagram of the traditional cement production process using a
long wet kiln.
Exhibit 13: Basic Cement Production Process
RAW
MATERIALS
GRINDING
MILL
_
Burning Zone
(Clinker Zone)
Calcining Zone
liiiiiiiiiiiiiijiiijijijijijiiiiiiiiiiiiiiiiii
Source: Report to Congress on Cement Kiln Dust (obtained from the EPA Sectors Notebook,
September, 1995).
Exhibit 14 shows the relative shares of the different fuel types used in the cement sector in
2002. Coal was the dominant fuel at about 58 percent of the total, not only because it was
relatively inexpensive but also because the resulting coal ash contains minerals that chemically
combine with the raw material to make clinker. The "Other" fuel category included petroleum
coke as well as waste materials that were incinerated for fuel, such as tire derived fuel (TDF).
Tires made up 3 percent of the total fuel use by this sector in 2004.44
44
See Table 22 in Energy Trends in Selected Manufacturing Sectors, EPA Sector Strategies Division,
March 2007.
29
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Exhibit 14: Primary Fuel Inputs as Fraction of Total Energy Supply for the Cement
Industry in 2002 (fuel use only)
Coal
58%
Other
23%
Net Electricity
11%
Natural Gas
5%
Coke & Breeze
2%
Source: Energy Trends in Selected Manufacturing Sectors, EPA Sector Strategies Division, March
2007, based on DOE/EIA Manufacturing Energy Consumption Survey, 2002 Data Tables. Totals may
not add to 100 percent due to rounding.
Net electricity provides a rough approximation for purchased power and includes purchased
electricity, transfers in, and generation from noncombustible renewables, and excludes electricity
transferred out or sold.
Air Emissions
The largest emission source within cement plants is the kiln operation, which includes the feed
system, preheater or precalciner (for preheater/precalciner kiln systems), the fuel firing
system(s), the actual kiln, and the clinker cooling and transport system. The kiln generates NOx,
SO2, CO, and hydrocarbon emissions. NOx emissions are from the combustion of fuels to dry,
calcine, and clinker the raw materials. The emissions of CO, SO2 and hydrocarbons are
primarily derived from organic and pyretic sulfur in the raw materials. Sources of particulate
emissions include raw material storage, grinding and blending, clinker production, finish
grinding, and packaging.
b. Location
There were 114 cement plants located in 37 states in 2005. Regionally, cement production was
concentrated in six states - California, Texas, Pennsylvania, Michigan, Missouri, and Alabama -
which accounted for approximately one-half of U.S. production.45 Exhibit 15 below presents the
location of cement plants in relation to nonattainment and maintenance areas for the current 8-
hour ozone NAAQS, as of October 2007.46 47 Exhibit 16 presents the location of cement plants
in relation to counties violating the current 24-hour PM2.5 NAAQS, based on 2004-06
monitoring data.48 While these exhibits indicate that a number of cement plants are located in
or near areas that currently violate the short-term standards for ozone or PM2.5, it should be
noted that many of these areas are projected to attain the standards in the near term while other
areas are at risk of violating the standards in the future. It should also be emphasized that
source-specific factors, including process-level emissions, stack parameters, topography,
meteorology and air chemistry, would need to be taken into account when assessing the air
quality impact of any individual facility or sector on local or regional air quality problems.
45 See Cement Profile in Sector Strategies Performance Report 2006, U.S. EPA Sector Strategies
Division, available atwww.epa.gov/sectors/performance.html.
46 The locations of cement plants were obtained from PIS Plant Directory, U.S. Cement Plants, December
2002. Data on nonattainment and maintenance areas for the current 8-hour ozone NAAQS were
obtained from EPA's Green Book: Nonattainment Status for Each County by Year, October 2007,
available at http://www.epa.gov/oar/oaqps/greenbk/map8hrnm.html and
http://www.epa.gov/oar/oaqps/greenbk/anay.html.
47 EPA proposed tightening the 8-hour ozone NAAQS on July 11, 2007. See:
http://www.epa.gov/air/ozonepollution/pdfs/20070711_proposal_fr.pdf.
48 The locations of cement plants were obtained from PIS Plant Directory, U.S. Cement Plants, December
2002. Data on counties violating the current 24-hour PM2.5 NAAQS were obtained from EPA's Air
Quality System (AQS) as of July 11, 2007, available at http://www.epa.gov/airtrends/values.html.
30
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
Exhibit 15: Location of Cement Plants in Relation to
8-Hour Ozone Nonattainment and Maintenance Areas
LEGEND
• Cement Plants
6-Hour Ozone Nonattainmenl Areas
a-Hour Ozone Maintenance Areas
Note: Partial nonallammenl counties u •-•. those v*lh pan
i_>1 the! county designated as non attainment artd part as
attainment) ate sho^-ii as nonattfnnment on the map.
Sources: Cement plant locations obtained from PIS Plant Directory, U.S. Cement Plants, December
2002. Nonattainment and maintenance areas for the current 8-hour ozone NAAQS obtained from
EPA's Green Book: Nonattainment Status for Each County by Year, October 2007.
Exhibit 16: Location of Cement Plants in Relation to
Counties Violating the 24-Hour PM2.5 NAAQS
LEGEND
• Cement Plants
Dl Counties Violating the 24-hour PM2 5 NAAQS
Note A oumbeT of *i<)ijrticm«l atuni.es. for which complete monrtotlng
data were not yet avauabte. may .afee, be dasignBted as violating the 24-
hout PM2 5 MAAGS (Mice 1he monitoring dr,M are oon^tete
Sources: Cement plant locations obtained from PIS Plant Directory, U.S. Cement Plants, December
2002. Counties violating the current 24-hour PM2.5 NAAQS (based on 2004-06 monitoring data)
obtained from EPA's Air Quality System (AQS) as of July 11, 2007.
31
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
c. Potential Episodic Control Measures
We focused our analysis on four categories of potential episodic control measures for the
cement sector:
» Fuel switching;
« Low-cost retrofits and enhanced use of existing control equipment;
« Scheduling changes; and
« Combustion reoptimization.
/'. Fuel switching
We analyzed three potential short-term fuel switching opportunities: (a) increasing the use of tire
derived fuel (TDF); (b) partially replacing coal with used oil, waste solvents or other alternative
liquid fuels; and (c) partially replacing coal with natural gas.
Increasing TDF usage
According to industry experts, burning TDF at cement kilns by means of mid-kiln injection can
reduce NOx emissions by as much as 20 to 40 percent.49 Moreover, TDF can be stored easily
by most cement plants in closed trailers onsite. Hence, we considered whether there was any
potential to increase TDF usage rates on an episodic basis to reduce NOx emissions and help
communities meet their ozone and/or PM2.5 attainment goals.
Currently, about 25 percent of the kilns in the U.S. are configured to use TDF as an alternative
fuel source. These kilns typically burn anywhere between 10 and 20 percent of their total fuel in
TDF, with the upper limit determined through existing permit requirements. Most of these kilns,
however, appear to use TDF below their maximum permitted levels, as well as below their
physical TDF burning capacities, due to cost and availability considerations. Hence, these kilns
have some potential for increasing their TDF feed rates.
Most industry analysts agreed, however, that because of the complex chemistry of the kiln
operation, making short-term changes to the TDF feed rates could have a negative impact on
the clinker chemical composition and may require other changes in the production process. In
addition, kilns typically operate under "steady-state" conditions, and short-term deviations due to
changes in the fuel mix can make them less efficient, thereby requiring more energy to operate
at the optimum level. Short-term changes in the fuel mix can also cause kiln upsets. Additional
energy consumption or a kiln upset could result in a corresponding increase in air emissions
which could undermine or erode the emission reduction benefits associated with greater TDF
usage.
Moreover, some analysts believed that the NOx emission reduction resulting from each
incremental unit of TDF burned in a kiln is likely to be less than the reduction that resulted from
the previous unit of TDF burned (that is, the relationship between increased use of TDF and its
associated emission reduction benefit may be nonlinear with diminishing returns). Increasing
TDF burn rates from current levels could, therefore, give rise to less than proportional
49 See also SCAQMD Expects Tires to Reduce NOx Emissions, available at
www.wbcsd.ch/web/projects/cement/tf2/tires.pdf.
32
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
reductions in NOx emissions, making the additional TDF consumption less useful as an episodic
pollution control strategy.50
Switching from coal to used oil or solvents
Another short-term fuel switching option that we investigated involves the introduction of a
limited quantity of used oil, waste solvent or other alternative liquid fuel to replace existing coal
used in a kiln.
Most analysts thought it might be possible to replace 5 to 15 percent of the coal burned in a kiln
with used oil without significantly altering the steady-state chemistry in the kiln. A switch from
coal to used oil of this magnitude could potentially produce significant incremental NOx
reductions - particularly if the first displaced units of coal yield a disproportionate NOx emission
reduction benefit. A fuel switch of this magnitude could also yield some incremental SO2
emission reductions. Similar emission reductions may also be achievable by displacing a
limited amount of the coal burned in a kiln with waste solvents or tar. The introduction of waste
solvents, tar or other alternative liquid fuels could, however, trigger the need for lengthy and
contentious permit revisions, and it could be difficult to justify such revisions where the fuels in
question would only be used for short intervals. For this reason, used oil may be a better
candidate for fuel switching as part of an episodic control strategy, particularly in cases where a
cement plant is already set up and permitted to burn used oil and has one or more dormant oil
tanks.
Switching from coal to natural gas
A final short-term fuel switching option we considered was to partially replace the coal burned in
precalciners with natural gas. According to industry analysts, typically 50-60 percent of the fuel
consumed by cement plants with precalciner kilns is burned in the precalciner, and coal is the
predominant fuel utilized. Unlike the kiln itself, where switching from coal to natural gas would
likely increase NOx emissions, displacing some of the coal burned in the precalciner with
natural gas would be expected to yield NOx emission reductions. However, the coal used in
different precalciners may contain different amounts of nitrogen (generating NOx after
combustion), and so this fuel switching option may be of limited utility as a NOx reduction
strategy at facilities that already burn coal with relatively low nitrogen content.51
A major concern expressed by industry analysts regarding this fuel switching strategy was that
replacing a large percentage of the coal burned in a precalciner with natural gas could alter the
chemical composition of the clinker (e.g., by reducing the coal ash content), thereby raising
quality concerns regarding the final product. It may, however, be possible to replace coal with
natural gas on a more limited basis without raising product quality concerns. This requires
further analysis.
50 One option that appears to be worth considering is to increase the number of kilns that are permitted to
burn TDF. Technically, this seemed to be a viable option, especially for the preheater/precalciner kilns,
most of which are currently not permitted to burn TDF. While permitting more kilns to burn TDF cannot be
considered an episodic control strategy, this strategy may hold promise for the ultimate goal of pollution
reduction in nonattainment areas.
51 An alternative to replacing coal with natural gas could be to replace coal with liquid petroleum gas
(LPG), which would be easier to store onsite than natural gas. While most analysts we spoke with did not
raise any immediate concerns with this type of fuel switch, further studies need to be conducted to
determine its viability.
33
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Industry analysts noted that given the high price of natural gas, switching from coal to natural
gas on a continuous basis would be prohibitively expensive. However, since this strategy
involves only short-term substitution of a limited portion of the coal used in the precalciner, and
since it would require very low capital expenditures, it may prove to be more cost-effective than
other available alternatives. However, site-specific costs, access to natural gas supplies and
other facility-specific factors would need to be taken into consideration in order to determine
whether this can be a viable episodic control strategy.
Summary: Fuel Switching
Overall, large scale fuel switching measures to reduce emissions from cement kilns on an episodic basis
seem to have little potential.
» The biggest concern expressed by industry analysts was the importance of maintaining the steady-
state nature of kiln operations which minimizes kiln upsets and related emission increases, and the
optimum fuel mix to drive kiln efficiency.
» A related concern was that the quality of the final product was dependent on the complex chemistry
inside the kiln; changing fuels on short notice may alter the chemical composition of the final product.
Our discussions with industry analysts did, however, indicate there may be potential to replace a limited
portion (e.g., 5 to 15 percent) of the coal burned in a cement kiln with used oil, and a limited portion of the
coal burned in a precalciner with natural gas. These relatively small scale fuel switches could potentially
yield additional SO2 and NOx reductions without disrupting the steady-state production process or the
chemistry in the kiln, although the magnitude of the pollution reductions would need to be analyzed to
determine their significance. These options require further analysis.
/'/'. Low-cost retrofits and enhanced use of existing control equipment
Another set of episodic control measures we considered entails the use of pollution control
equipment, such as post-combustion control technologies. Although most pollution control
retrofits require significant capital expenditure for installation, there are some that have relatively
low installation costs but higher O&M costs. Such low-capital, high O&M retrofits may be cost-
effective to operate for short time periods during an episodic event. Similarly, where control
technologies are already deployed, it may be possible to run them more aggressively for short
intervals, or in other ways that would not be feasible or cost-effective on a continuous basis.
« One control technology that seemed promising from the standpoint of achieving further
short-term NOx reductions was Selective Non-Catalytic Reduction (SNCR). 52
« Another promising control option from the standpoint of achieving further short-term SO2
reductions was the use of lime injection techniques, including micro-fine lime addition.
SNCR for NOx Control
Analysts felt that SNCR may hold some promise as an episodic control strategy for reducing
NOx emissions from preheater/precalciner kilns in the cement industry. SNCR technology has
been in use for many years. To date, however, fewer than ten cement plants have installed an
SNCR system. While there are indications that some state and local air quality regulators may
52 Another NOx control technology that might be viable for the cement sector is gas reburn. Industry
analysts pointed out there is no data on gas reburn in the cement sector. However, because gas reburn
may be a viable strategy for episodic NOx reductions in other sectors, it warrants further study with
respect to the cement sector as well.
34
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
be considering the imposition of SNCR control requirements where further NOx emission
reductions are needed to achieve air quality goals, the limiting factor to date has been the cost
of implementing these systems.
According to industry analysts and technology vendors, the cost of purchasing and installing an
SNCR system may be on the order of $1 million for a typical mid-sized kiln. However, for most
large applications such as cement kilns, the majority of the total annualized cost associated with
an SNCR system is for the reagent used (urea or aqueous ammonia). Therefore, the cost
burden associated with installing and operating an SNCR system could be reduced significantly
under an episodic control program. For example, it might be possible to reduce the total cost
burden of an SNCR system by 50 percent or more where the system is installed and run at an
intermediate control level on a continuous basis and then ramped up to full capacity on days
with high air pollution concentrations. Even greater cost reductions may be possible where the
system is kept on standby and used only on days with high air pollution concentrations.
A major concern expressed by industry analysts was that once an SNCR system is installed and
available for use as part of an episodic control program, regulators and/or community activists
may demand that it be run at full capacity on a continuous basis. Further analysis of the costs
of this option and other site-specific considerations is required.
For those cement plants that have already installed an SNCR system, an alternative strategy
may be to increase their utilization rates for short time periods when high air pollution
concentrations occur. Although we were not able to conclusively determine whether there is
room for increased utilization at facilities with existing SNCR systems, this strategy may warrant
further exploration with regulators, plant managers and equipment vendors.
Lime injection for SO2 Control
Cement plants may have the ability to use lime injection to get additional cost-effective SO2
removal on a short-term basis. Since most cement plants using a preheater or precalciner
system already have spray cooling and conditioning towers, the additional costs associated with
lime injection would be less than for wet or long dry kilns. Further research is needed to
determine the extent to which lime injection could achieve further SO2 reductions on an episodic
basis, and the cost of implementing this approach. Site-specific factors would need to be taken
into consideration as well.
Micro-fine lime addition for SO2 Control
Micro-fine lime addition involves the use of a very fine grind lime to absorb SO2 in the exhaust
stream of a kiln. The micro-fine lime is converted into a slurry (water suspension of lime) and
injected into the water that is used in a spray cooling and conditioning tower that is connected to
a bag house or ESP. The capital costs for retrofitting a cement plant to use micro-fine lime
addition could be relatively small where a facility already has a spray cooling and conditioning
tower in place, as is the case with most preheater/precalciner kilns. There would, however, be
some capital costs to purchase and install equipment needed to store the micro-fine lime
reagent onsite and for other equipment modifications. However, the major cost associated with
micro-fine lime addition is the relatively high cost of the reagent itself, which makes it unsuitable
for use as a continuous control strategy for cement kilns.
Evidence of micro-fine lime addition's use in the cement industry is limited, with its use to date
restricted to precalciner kilns. Its primary function has been either to serve as a backup SO2
control measure, e.g., to meet permit limits during short periods when the raw mill is not in
operation, or to address detached plume/plume opacity issues created by condensing ammonia
35
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
salts. When the raw mill is operating it serves as a de facto scrubber capable of removing as
much as 80 to 90 percent or more of the SO2. (Similar SO2 removal efficiencies are typical with
the operation of other types of kilns.) Nonetheless, our discussions with industry analysts
indicated there might be opportunities to use micro-fine lime addition to achieve additional SO2
reductions at precalciner kilns even when the raw mill is operating, as well as at other types of
kilns. According to its manufacturer, micro-fine lime addition could be expected to further reduce
SO2 emissions from cement plants by 50 to 75 percent. This technology should be evaluated
further to determine its viability and cost as an episodic control measure.
Summary: Low Cost Retrofits and Enhanced Use of Existing Control Equipment
Most analysts agreed that installing an SNCR system on preheater/precalciner kilns may have promise as
an episodic control strategy for reducing NOx emissions and warrants further research. Because of the
relatively low installation costs and the limited use of reagent, installing and using SNCR systems as part
of an episodic control program may be a cost-effective strategy for preheater/precalciner kilns in this
industry.
For those cement plants that have already installed an SNCR system, an alternative strategy may be to
increase their utilization rates for short time periods during an episodic event. Although we were not able
to conclusively determine whether there is room for increased utilization at facilities with existing SNCR
systems, this strategy may warrant further exploration.
The use of lime injection or micro-fine lime addition to achieve additional SO2 reductions at cement kilns
should also be evaluated further to determine their viability and costs as episodic control strategies.
/'/'/'. Scheduling changes
We considered two types of short-term scheduling changes for the cement industry:
« First, we considered short-term changes in the clinker production schedule that could be
undertaken to achieve additional emission reductions on a limited number of days when
high air pollution concentrations occur. An example of such a schedule change is the
curtailment of kiln operations for short periods.
» The second type of scheduling change relates to cement manufacturing activities that are
ancillary to, and independent of, the actual clinker production process.
Changes to the clinker production schedule
Industry analysts indicated that while kiln operations could be curtailed on short notice (e.g., by
slowing the rotation of the kiln to reduce its production rate), cement plants, which have been
operating near full capacity in recent years, may be unable to make up the foregone production.
Moreover, kilns that operate at optimum efficiency generally have the lowest emissions intensity
(i.e., emissions per unit of clinker produced). Therefore, short-term changes to the optimum
production rate can be counter-productive to emission reduction goals, eroding the short-term
emission reduction benefits resulting from a slowdown while increasing the plant's total annual
emissions.
Scheduling changes related to ancillary activities
Activities that are ancillary to the operation of a cement kiln include quarrying, running the
finishing mill and bagging/loading the finished product. In addition to their direct emissions
impact, these ancillary activities require electricity, and most cement kilns depend on grid-
supplied electricity for these activities. Therefore, episodic control measures that curtail these
36
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
operations for short periods of time can provide secondary emission reductions by temporarily
reducing demand for electricity.
» Defer quarrying activities: Quarrying activities at cement plants can potentially be deferred
to reduce PM and NOx emissions from diesel equipment, as well as the demand for grid-
supplied electricity, on days with high air pollution concentrations. According to industry
analysts, some facilities already conduct quarrying activities at night (for example, in
response to interruptible electricity contracts, or to take advantage of cheaper off-peak
electricity rates), and there may be additional opportunities to defer quarrying activities until
nighttime hours as an episodic control measure. Other facilities may be able to curtail
daytime quarrying activities when high air pollution concentrations occur by using existing
stockpiles. These relatively low-cost strategies could be particularly useful on high electricity
demand days when electric utilities may rely on relatively high emitting generating units in
order to meet peak-day demand. Curtailing peak-day electricity consumption, or shifting
that consumption from daytime to nighttime (when lower-emitting generating capacity is
available and the resultant emissions may have less ozone or PM2.5 forming impact), could
provide significant air quality benefits.
« Defer operation of finishing mill: The finishing mill uses the clinker generated by the kiln
to produce the finished product sold to buyers. Finishing mills require electricity, and there
is evidence that some mills run at night to take advantage of cheaper off-peak electricity
rates, and others operate under interruptible electricity contracts so that they can be
temporarily halted when there is a surge in electricity demand. Thus, some cement plants
may be able to defer running the finishing mill in order to reduce electricity demand when
high air pollution concentrations occur. As discussed above, this strategy could be
particularly useful on high electricity demand days when electric utilities may rely on very
high-emitting generating units in order to meet peak-day demand. The extent to which
operations at a finishing mill can be deferred depends, however, on the facility's storage
capacity and the extent to which the mill has excess grinding capacity to make up the lost
production.
» Defer bagging/loading: Another ancillary activity that we considered with regard to short-
term schedule changes relates to plant bagging and bulk loading processes. Changes in
bagging and bulk loading schedules have the potential to defer electrical demand to off-peak
hours, along with emissions from the operation of diesel trucks and equipment. However,
according to industry analysts, many cement plants are set up to provide round-the-clock
cement loading services to their customers. Since most customers rely on their own trucks
and may not have the flexibility to defer deliveries, these plants need to load vehicles on a
continuous basis as they arrive. In other cases, loading at night may create noise related
problems with trucks traveling through residential neighborhoods. For these reasons,
scheduling changes pertaining to bulk loading operations do not appear to hold promise as
episodic control measures. However, industry analysts indicated that most facilities have
some inventory of bagged product, and this may allow for short-term deferral of bagging
operations on days with high air pollution concentrations. Nonetheless, there are indications
that bagging operations account for only a small portion of a cement plant's overall
emissions.
37
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Summary: Scheduling Changes
Our discussions with industry analysts indicated there may be opportunities to implement short-term
scheduling changes in certain ancillary activities without significant disruptions to kiln operations or facility
production schedules. For example, some cement plants may be able to defer quarrying activities and/or
operation of the finishing mill on days with high air pollution concentrations. Most plants already have the
infrastructure to stockpile some quantity of raw materials or finished products. Thus, rescheduling these
ancillary activities for limited time periods may be feasible, if given up to 24-hours advance notice.
iv. Combustion reoptimization
We also considered the following potential short-term combustion reoptimization techniques
with regard to their ability to achieve incremental NOx emission reductions from the cement
industry on days when high air pollution concentrations occur:
» Kiln operational control adjustments (such as controlling flame temperature or oxygen
content).
» Cement kiln dust (CKD) recirculation and/or flue gas recirculation.
Combustion reoptimization techniques seem to hold little promise as an episodic control
strategy for the cement sector. Most cement kilns, particularly in ozone nonattainment areas,
already use sophisticated computerized control systems with combustion controls optimized for
maximum kiln efficiency and product quality. And because cement kilns are heavily dependent
on their steady-state performance for optimum operational efficiency, most analysts suggested
there would be little room to reoptimize these systems on a short-term basis to obtain significant
emission reduction benefits. Any short-term changes may actually reduce kiln efficiency and
increase the emissions intensity of kiln operations.53
Summary: Combustion Reoptimization
Industry analysts believe that most kilns are currently optimized; any short-term modifications would likely
reduce operational efficiency and could potentially increase emissions.
53 One area that may need further evaluation is to determine whether there is a tradeoff potential in
emission reduction goals through combustion reoptimization. For example, there appear to be ways to
reoptimize kiln operations to further reduce NOx emissions for short intervals when high concentrations of
ozone and PM2.5 occur at the cost of increasing CO emissions during those time periods.
38
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
d. Opportunity Assessment of Potential Episodic Control Options
Exhibit 17 summarizes the potential viability of the four primary options for further reducing
emissions on days with high air pollution concentrations.
Exhibit 17: Opportunity Assessment - Cement
Option
Fuel switching
Low cost retrofits and
enhanced use of existing
control equipment
Scheduling changes
Combustion reoptimization
Viability as Possible Episodic Control Measure
Unlikely- Fuel switching on a large scale
Promising - Small scale displacement of coal with other fuels in the kiln
or precalciner (for NOx reductions)
Promising - SNCR for preheater/precalciner kilns (for NOx reductions)
Possible - Micro-fine lime addition for facilities with spray cooling and
conditioning towers (for SO2 reductions)
Promising - For ancillary activities, such as deferring quarrying and/or
running the finishing mill (to reduce direct emissions and electricity
demand)
Unlikely
39
-------
EPISODIC AIR POLLUTION CONTROL MEASURES
JANUARY 2008
IV. Implications for Other Sectors
Although our primary focus in this study has been on three selected industry sectors-pulp and
paper, iron and steel, and cement-some of the episodic control measures explored here may be
applicable to other sectors as well. This section of the report assesses the implications of our
findings for emission sources in other industry sectors.
a. Implications for Industrial Boilers
Episodic control measures applicable to industrial boilers in the pulp and paper sector, or iron
and steel sector, may also be applicable to boilers in other manufacturing sectors. Exhibit 18
below shows the distribution of industrial boilers, by size, across major manufacturing sectors.
Exhibit 18: Industrial Boiler Inventory - Boiler Capacity
(Number of Boiler Units in Parenthesis)
Boiler
Capacity
(MMBtu/
hr input)
<10
10-50
50-100
100-250
>250
Total
Food
31,070
(6570)
64,970
(3,070)
37,885
(570)
47,950
(330)
27,860
(70)
209,735
(10,610)
Pulp&
Paper
4,105
(820)
24,490
(1 ,080)
36,665
(530)
81,500
(540)
229,590
(490)
376,350
(3,460)
Chemicals
28,660
(6,720)
81,690
(3,370)
64,970
(950)
86,840
(590)
150,915
(350)
413,075
(11,980)
Petroleum
Refining
1,255
(260)
6,670
(260)
18,390
(260)
30,480
(200)
114,720
(220)
171,515
(1,200)
Metals
7,505
(1 ,850)
19,405
(920)
22,585
(330)
17,775
(110)
45,365
(120)
112,635
(3,330)
Other
Manufactur
ing
29,710
(7,275)
80,585
(3.680)
62,630
(930)
62,790
(440)
47,760
(110)
283,475
(12,435)
Total
102,305
(23,495)
277,810
(12,380)
243,125
(3,570)
327,335
(2,210)
616,210
(1,360)
1,566,785
(43,015)
Source: ICF/EEA Boiler study for ORNL.
As is evident from Exhibit 18, the chemical industry accounts for more than a quarter of the
boilers used in all manufacturing sectors, in terms of total capacity, total number of units, and
total number of large units (i.e., over 250 MMBtu/hr input). Use of industrial boilers in the food
industry (consisting of 13 percent of total capacity, 25 percent of the total number of units, and 5
percent of the total number of large units) and petroleum refining industry (consisting of 11
percent of total capacity, 3 percent of the total number of units and 16 percent of the total
number of large units), also make these industries particularly important candidates for the
40
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
consideration of episodic control strategies where such strategies can help address air quality
problems.
While there are sector-specific (and indeed, facility-specific) issues to analyze further, this
section provides a general overview of a number of control measures that might have broad
applicability to industrial boilers. Each of these potential measures is discussed in more detail in
one or more of the preceding sections, including a more complete discussion of potential
limitations and concerns.
/'. SO2 control measures
Our discussions with regulatory experts and industry analysts indicated there is a large
population of industrial boilers that do not currently employ any SO2 controls technologies.
Boilers, including large boilers that pre-date the New Source Performance Standards for
industrial boilers that came into effect in 1986, typically have no federal requirements for SO2
control (although they may be subject to state-level requirements or consent decrees that
require the use of control technology or low sulfur coal).
In situations where further SO2 reductions are needed to address short-term air quality problems
(e.g., violations of the 24-hour PM2.5 standard), but where installing a packed tower scrubber or
other continuous SO2 control device is not technically or economically feasible, facilities may be
able to install a lime injection system to temporarily achieve additional SO2 emission reductions
on short notice as part of an episodic control strategy. For example, boilers that use a
baghouse or ESP for PM control could intermittently inject lime powder into the duct that carries
exhaust gases from the boiler to the baghouse or ESP to obtain further SO2 reductions on a
short-term basis. Boilers that do not have post-combustion SO2 controls may also be able to
reduce SO2 emissions by switching to natural gas or a low-sulfur coal for limited periods of time,
where those fuels are not currently being utilized.
In addition, industrial boilers that have been retrofitted with SO2 control devices may have the
ability to increase the control efficiencies of those devices, or to switch to a lower sulfur fuel, for
short periods of time in certain situations.
/'/'. NOx control measures
SNCR systems may hold some promise as a short-term episodic control measure for reducing
NOx emissions from industrial boilers in many industry sectors. SNCR technology has been in
use for many years. However, in most sectors, its use has been limited to boilers that are
subject to new source control requirements. While there are indications that some state and
local air quality regulators may be considering the imposition of SNCR control requirements on
existing industrial boilers in some industries, in situations where further emission reductions are
needed to achieve air quality goals, the limiting factor to date has been the cost of installing and
operating this technology.
According to industry analysts and technology vendors, the cost of purchasing and installing an
SNCR system may be on the order of $1 million for a large industrial boiler. However, for most
large boiler applications, the majority of the total annualized cost associated with an SNCR
system may be for the reagent used (urea or aqueous ammonia). Therefore, the cost burden
associated with installing and operating an SNCR system could be reduced significantly under
an episodic control program. For example, in some situations it might be possible to reduce the
total cost burden of an SNCR system by 50 percent or more if the system were installed and run
41
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
at an intermediate control level on a continuous basis, and then ramped up to full capacity on
days with high air pollution concentrations. The cost burden could be further reduced if the
SNCR system were kept on standby and used only on days with high air pollution
concentrations.
Another NOx control technology that might be used as part of an episodic control program is
gas reburn. According to industry analysts and EPA experts, many industrial boilers have
access to natural gas for use as a backup or starter fuel. For these boilers, limited capital
expenditures would be required to install additional burner(s) and duct work to tie the burners to
the gas supply. The majority of the cost would be for the additional natural gas needed to
operate the technology. Because under an episodic control program the reburn technology
would be used only on a limited number of days each year, the system may not be cost
prohibitive even under the current scenario of high gas prices.
Finally, some industrial boilers may also have the ability to switch to fuels with lower nitrogen
content or flame temperatures for short periods of time to achieve additional NOx emission
reductions when high air pollution concentrations occur.
/'/'/'. PM control measures
Because of the history of regulatory requirements for PM abatement, the vast majority of
industrial boilers appear to be controlled for primary PM. Most large boilers currently use
baghouses, ESPs or venturi scrubbers for PM control purposes. While the industry analysts we
spoke with were uncertain about how much opportunity there may be to increase the control
efficiencies of these devices on a short-term basis, they felt there might be some situations in
which this could be accomplished. For example, in some cases it may be possible to increase
the voltage in an ESP in order to increase its PM control efficiency for short time periods. In
addition, some boilers may be able to further reduce their PM emissions on a short-term basis
by switching to a lower emitting fuel.
b. Implications for Other Industrial Sources
As with industrial boilers, some of the episodic control measures applicable to other units and
processes within the three selected industries may be applicable to similar units or processes in
other industries as well. Examples include fuel switching at furnaces and process heaters,
installation of low capital cost retrofits, enhanced use of existing control equipment, and certain
scheduling changes. Some sectors may also be able to provide additional, sector-specific
opportunities for reducing emissions on an episodic basis. For example, in the course of our
research we learned that a number of chemical facilities in the Gulf Coast region defer routine
maintenance activities on high ozone days as part of local ozone action programs. Perhaps the
most significant of the deferred maintenance activities relates to planned shutdowns. When
shutting down a chemical plant for routine maintenance, facility operators typically need to flare
off some of the VOCs from process vessels, and this can be a major short-term emission
source.54 Hence, the deferral of planned shutdowns at chemical facilities on high ozone days
can avoid emission spikes at times when they would be most detrimental.
54 A recent study of the Houston-Galveston area suggests that flaring is a major source of highly reactive
VOCs from point sources in the area and, therefore, a potential major contributor to violations of the 8-
hour ambient air quality standard for ozone. See Murphy, C.F. and D. T. Allen, "Hydrocarbon Emissions
from Industrial Release Events in the Houston-Galveston Area and their Impact on Ozone Formation,"
Atmospheric Environment, 39, 3785-3798 (2005).
42
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
c. Implications for Electric Utilities
Several episodic control measures explored in this analysis involve the rescheduling of
industrial processes that use large amounts of grid-supplied electricity (e.g., quarrying activities
and finishing mill operations at cement plants, and melting operations at steel mini mills), to
times when they are less likely to exacerbate air quality problems. Other industry sectors may
have similar opportunities to temporarily halt, curtail or defer processes that use large amounts
of grid-supplied electricity, as part of an episodic control program. These measures could be
particularly useful on high electricity demand days when electric utilities may rely on very high
emitting generating units in order to meet peak electricity demand. Reducing electricity demand
from industrial sources on high electricity demand days, or shifting that demand from daytime to
nighttime hours (when lower-emitting generating capacity is available and the resultant
emissions may have less ozone or PM2.5 forming impact), could provide significant air quality
benefits.
In addition to taking advantage of efforts by industrial sources to reschedule electricity-intensive
operations, electric utilities may have the ability to employ other episodic control measures to
reduce NOx, SO2 and/or PM emissions on days when high air pollution concentrations occur. A
number of such measures may be possible and warrant further research, including fuel
switching strategies, dispatching strategies, installation of low capital cost retrofits, and
enhanced use of existing control equipment.
43
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
3. Summary of Findings & Next Steps
As EPA, states, communities and other stakeholders consider the potential role that episodic
control measures could play in attaining and maintaining air quality standards for ozone and
PM2.5, one challenge will be to determine what measures can be viably and cost-effectively
implemented by industrial sources on an episodic basis. This analysis identifies and
preliminarily assesses a potential set of episodic control measures based on a review of
literature and discussions with industry experts. In order to put some bounds on the research,
we choose three industry sectors to serve as case studies for this analysis - pulp and paper,
iron and steel, and cement. These sectors, however, are in no way an exhaustive list of sectors
where episodic control measures may potentially be employed. Indeed, a number episodic
control measures were identified in the course of this research that could have broad application
to emission sources in other industries.
I. Findings
a. Episodic Control Measures Exist for the Three Selected Sectors
This study finds that there are a number of potential episodic control measures available for
each of the three sectors analyzed. Further exploration of these measures is warranted to
determine the extent to which they would provide net emission reductions, and to more fully
assess the technical, cost, and operational impacts of the various measures on individual
facilities.
» All three sectors possess some fuel switching capabilities and may be able to shift to lower
emitting fuels to reduce NOx, SO2, and/or PM on an intermittent basis for short periods of
time. The manner and extent to which short-term fuel switching could be employed,
however, varies from sector to sector. Since facilities typically select their fuels so as to
minimize production costs and meet other operational objectives, switching to more
expensive fuels on a continuous basis would be untenable for many facilities. However,
because fuel switching under an episodic control program would only be invoked on a
limited number of days each year, the total incremental costs may be within acceptable
levels for many facilities.
» Installation of low capital cost retrofit technologies, such as selective non-catalytic reduction
(SNCR) and gas reburn (for NOx control), and lime injection (for SO2 control) appears to
hold promise in the context of an episodic control program. Although the high O&M costs
associated with these retrofit technologies (e.g., due to the significant consumption of
reagent or natural gas) may make them cost prohibitive for use on a continuous basis at
many facilities, they may be suitable for use at some facilities as an episodic control
measure.
» Opportunities may exist to modify - and thereby enhance - the use of existing pollution
control equipment on a temporary basis to control for additional pollutants (i.e., in addition to
the main pollutant(s) these controls are intended for). For example, many industrial boilers
and furnaces have venturi scrubbers. These scrubbers are generally installed as PM control
devices but may also remove some SO2 from the exhaust stream of the combustion unit.
Some industry analysts and regulatory experts believe there may be an opportunity to
enhance the SO2 removal efficiency of venturi scrubbers in some situations by adding alkali
44
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
reagent. This may be technically feasible and economical where a unit already has such a
device and is able to obtain significant incremental SO2 reductions on a short-term basis
with the addition of the reagent.
« Opportunities may also exist to run existing pollution control devices more aggressively for
short intervals to achieve incremental emission reductions. Some industry analysts thought
that in some cases it might be possible to temporarily increase the control efficiency of a
variety of NOx, SO2, and PM control systems on days when high air pollution concentrations
occur and where, for cost or other reasons, existing permit conditions do not require these
control devices to be utilized at their maximum performance levels.
» While temporary shutdowns and curtailments that could disrupt core production processes
were generally deemed inadvisable, there may be opportunities to curtail certain operations
or rearrange production schedules on a short-term basis to move high emitting processes to
times when they are less likely to exacerbate air quality problems. For example, because
steel mini mill operation is generally batch in nature, some mini mills may be able to
reschedule melting operations from daytime to nighttime hours to reduce VOC and NOx
emissions, as well as electrical demand, during critical daylight hours when ozone formation
is likely to occur. There may also be opportunities to defer certain ancillary activities at
cement plants, such as quarrying activities and finishing mill operations, to reduce emissions
as well as electrical demand on a short-term basis.
» Changing the dispatch of boilers at large pulp and paper mills with excess boiler capacity
appears to be a feasible episodic control strategy in some instances.
» Additionally, there may also be other sector-specific episodic control opportunities, such as
increasing the use of cogeneration units in the pulp and paper sector, which require further
exploration.
» Episodic control measures related to combustion reoptimization techniques appeared less
viable for the selected industry sectors. Most facilities in these sectors use sophisticated
computer controlled equipment on highly inter-connected and synchronized production
processes. Industry experts seemed reluctant to entertain the idea of altering the already
optimized processes, as they might make the processes less efficient and/or more polluting.
b. Episodic Control Measures Also Exist for Other Industrial Sources
While this study has focused primarily on facilities in the pulp and paper, iron and steel, and
cement manufacturing industries, our research and information gathering indicates that a wide
variety of episodic control measures may be available for use by other industrial sources as
well.
» A number of the episodic control measures that are applicable to industrial boilers in the
pulp and paper sector, or the iron and steel sector, may also be broadly applicable to boilers
in other industry sectors. For example, some industrial boilers in other industries may be
able to install low capital cost retrofit technologies (such as SNCR or gas reburn for NOx
control, or lime injection for SO2 control), to reduce boiler emissions on a short-term basis.
Other boilers may be able to temporarily reduce emissions by switching to a lower emitting
fuel for limited periods of time. In addition, some boilers that have already been retrofitted
with control devices may have the ability to increase the control efficiencies of those devices
45
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
by modifying their operation (e.g., adding reagent to a venturi scrubber) or running them
more aggressively for short time intervals (e.g., increasing voltage in an ESP).
» As with industrial boilers, a number of the episodic control measures that are applicable to
other units and industrial processes in the three selected industries may be applicable to
similar units or processes in other industries as well. Examples include fuel switching at
furnaces and process heaters, installation of low capital cost retrofits, enhanced use of
existing control equipment, and certain scheduling changes. Some sectors may also be
able to provide additional, sector-specific opportunities for achieving emission reductions on
an episodic basis.
» Several of the episodic control measures examined in this report involve the rescheduling of
electricity-intensive industrial processes to times when they are less likely to exacerbate air
quality problems. Other industry sectors may have similar opportunities to temporarily halt,
curtail or defer electricity-intensive processes as part of an episodic control program. These
measures could be especially helpful on high electricity demand days when electric utilities
may rely on very high emitting generating units in order to meet peak-day demand for
electricity. Reducing peak-day electricity consumption, or shifting that consumption from
daytime to nighttime hours (when lower-emitting generating capacity is available and the
resultant emissions may have less ozone or PM2.5 forming impact), could provide
significant air quality benefits. Electric utilities may also have the ability to employ a variety
of other episodic control measures to reduce NOx, SO2 and/or PM emissions on days with
high air pollution concentrations.
Industry analysts and regulatory experts emphasized that because of the inherent differences
between facilities within a sector (as well as between sectors), generalized solutions for all
facilities within a sector are unlikely to be effective. Episodic control measures need to be
assessed on a plant-by-plant basis to determine their technical feasibility, practicality and cost-
effectiveness.
II. Next Steps
This report is intended to provide a preliminary assessment of the wide array of episodic control
measures that may be available to industrial sources. The potential measures identified in this
report should be studied further as EPA, states, communities, and other stakeholders consider
the potential role that episodic control measures could play in helping to meet ozone and PM2.5
air quality goals. A number of refinements to this analysis and additional research activities are
recommended below to help EPA and other stakeholders more fully evaluate the potential
contribution of episodic control measures.
» This analysis focuses primarily on three industry sectors - pulp and paper, iron and steel,
and cement. While it also provides a cursory assessment of the potential applicability of
episodic control measures to industrial sources outside of these sectors, further analysis
with respect to other industry sectors is needed.
» To better assess the technical feasibility, emission reduction potential and cost of
prospective episodic control measures, it will be necessary to conduct detailed engineering
and cost analyses of specific measures. Because of the heterogeneity across facilities
within a sector, as well as across sectors, these analyses will need to take into account site-
46
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
specific and process-level emissions, as well as other source-specific considerations at
individual facilities.
In order to more fully assess the business and operational impacts of potential episodic
control measures, it will be necessary to hold additional discussions with industry
representatives and technical consultants.
Because the ambient impact of episodic control measures is likely to be spatially
differentiated, location-specific air quality modeling will be necessary to determine the actual
contribution that short-term emission reductions from any particular industrial source or
sector would make in solving local and/or regional air quality problems.
In order to evaluate the potential utility of episodic control measures it will be necessary to
review state-of-the-art practices pertaining to the prediction of ozone and PM2.5 episodes.
The effectiveness of episodic control measures would in part depend on the degree of prior
notice that can be provided to plant operators, which is in turn a function of the ability of
regulatory agencies and/or other institutions to predict air pollution episodes. Certain
strategies may be technologically feasible and cost-effective, but may not be good
candidates to implement if adequate prior notice cannot be provided to the affected facilities.
Another key step in the further exploration of episodic control strategies is to quantify the
cost-effectiveness of these strategies on a dollar-per-ton of pollutant removed basis. Our
discussions with industry analysts indicated that this is a major unknown, because episodic
strategies for industrial sources have traditionally been overlooked in the air pollution control
literature. It must be emphasized, however, that comparisons to cost-per-ton estimates
based on continuous control programs would provide an erroneous picture, as the
acceptable cost-effectiveness range for an episodic control program is likely to be
substantially higher than the corresponding range for a continuous program. Thus, a better
understanding of the range of cost-effectiveness values that could be appropriate given the
benefits of an episodic control program needs to be carefully developed.
A number of critical policy questions will also need to be addressed before episodic control
measures can be systematically implemented. For example, how should these measures
be integrated with continuous and seasonal control measures as part of an air quality
management plan, and how might this vary based on the attainment status of an area?
Should episodic control measures be mandated by regulation, negotiated as part of the
permitting process, or pursued by individual facilities on a purely voluntary basis? What
types of monitoring, record-keeping and reporting provisions should be required to ensure
that episodic emission reductions are quantifiable and/or enforceable?
47
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
Appendix:
Examples of Reactions by Industrial Sectors to the
California Outages of 2001 and 2005
This Appendix discusses the challenges associated with unscheduled shutdowns in the pulp
and paper, iron and steel, and petroleum refining industries due to the rolling blackouts that
occurred in California in 2001 and 2005.
In large plants including pulp, paperboard, and paper mills, iron and steel mills, and petroleum
refineries, the operations of their major processes and ancillary equipment are intertwined and
synchronized. An unscheduled stoppage of a process or equipment could cost substantial
revenue losses to a plant because it not only stops the operation of the process or equipment,
but could also cause the breakdown of other units linked to the original unit. As such, the
reliability of each process or piece of equipment has to be high. To maintain the reliability of
equipment or entire production operations, industrial plants schedule regular maintenance of
their equipment and processes. Some of these maintenance activities could require a stoppage
of the entire plant's operation or the shutdown of particular units or groups of units in the plant.
The stoppage schedule is designed as far in advance as possible to ensure worker safety,
sufficient labor resources, minimal revenue loss, and equipment integrity upon the resumption of
plant/equipment operations.
During the rolling blackouts in California in 2001 and again in 2005, manufacturing facilities
were adamant that they should be informed well in advance and even asked for exemptions
from blackouts. Industries were concerned that blackouts could be extremely costly (due to loss
of revenue) and could also compromise worker safety. Petroleum refiners in California cited
that refiners could not resume operations with just a simple flick of a switch. Petroleum
refineries claimed that an hour-long shutdown could lead to a 2-3 day restart since tanks must
be drained and heating units must be brought back to temperature. Also, refiners already try to
follow a well-choreographed schedule of maintenance outages to ensure optimal operations in
their plants, so that an unscheduled stoppage could put the maintenance schedule in disarray.
A possible unacceptable result of an unscheduled refinery shutdown or curtailment is the loss of
gasoline/diesel fuel supply in the market, and as such could lead to an increase in
gasoline/diesel fuel prices. Historically, refinery outages have generally been scheduled when
markets are not tight (during non-summer season), and they therefore have had little or no
measurable impact on monthly average prices. However, a government program that requires
refiners to shutdown during the summer ozone season associated with peak gasoline demand,
could result in a loss of gasoline/diesel supply in the market and a corresponding increase in
fuel prices.
For the iron and steel industry, the implementation of a stoppage should differentiate between
the integrated mill and the mini mill. An integrated steel mill is comparable to a petroleum
refinery in the complexity of its processes and equipment. As such, an unplanned stoppage
could also cause substantial costs apart from risks to worker safety. For instance, an
unplanned shutdown of the blast furnace results not only in production losses but risks making
the blast furnace unusable because of the iron solidifying in the blast furnace' air tuyeres. The
cost to rebuild a blast furnace is estimated at approximately $100 million.
For mini mills, which use electric arc furnaces, it is possible to consider not a stoppage but a
rescheduling of operations. During the California rolling blackouts, some mini mills with
48
-------
EPISODIC AIR POLLUTION CONTROL MEASURES JANUARY 2008
interruptible power contracts switched to firm contracts with their power suppliers because the
power curtailments were too costly. Others rescheduled their melting operations to off-peak
hours to enjoy the cheaper electricity rates during those times and still allow for the interruptible
contract to continue. Note that a rescheduling to off-peak hours for the mini mills is a possibility
because the operation is generally batch in nature.
Pulp, paper, and paperboard mills, like integrated steel mills and petroleum refiners, have a
complex system of continuous operations that are well-synchronized. An unscheduled
stoppage could be very expensive for the company. Paper mills have many large boilers and
their start-ups require several hours. Several hours of lost production could be costly for a
paper company. It could also cost the reduction of equipment reliability and lower product
quality.
Paper mills, integrated steel mills, and refiners may also be large cogenerators and a number of
them are Qualifying Facilities (QF). QFs are essential to the balancing of the power grid and so
a shutdown of an industrial QF could cause demand-supply imbalance as well.
Another commonality among these industries is that they are capital-intensive so that
maximizing the utilization of their assets is essential to their profitability. Also, these companies
have substantial contracts with labor, their suppliers and their customers, and unmet obligations
due to unscheduled stoppages of their operations could also have legal implications.
49
------- |