vvEPA
United States
Environmental Protection
Agency
UNIVERSAL INDUSTRIAL SECTORS
INTEGRATED SOLUTIONS MODEL FOR PULP
AND PAPER MANUFACTURING INDUSTRY-
UNIVERSAL ISIS-PNP
i

-------
Universal Industrial Sectors Integrated Solutions
Model for the Pulp and Paper Manufacturing
Industry-Universal ISIS-PNP
November 2014
US Environmental Protection Agency
Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, North Carolina 27711
ii

-------
Contact Infor	mation
This document was prepared by staff from the Office of Research and Development and Office of Air Quality
Planning and Standards, US Environmental Protection Agency. Questions related to this document should be
addressed to Dr. Gurbakhash Bhander, US Environmental Protection Agency, Office of Research and
Development, Research Triangle Park, North Carolina 27711 (email: bhander.gurbakhash(5)epa.gov).
Acknowledgments
Under USEPA Contracts EP-C-09-027 and EP-C-09-029, Dr. James Staudt of AndoverTechnologies developed the
technical information on controls and energy efficiency measures used in the Universal ISIS Model for the Pulp
and Paper Manufacturing Industry. Dr. Ravi Srivastava, Ms. Elineth Torres, Mr. Charles Fulcher, Dr. Alex
Macpherson, Mr. William Yelverton, Mr. John Bradfield, Dr. Kelley Spence, Dr. Nabanita Modak and Mr. Amit
Srivastava of EPA provided technical support and many quality assurance checks. Finally, Dr. Wojciech Jozewicz,
Dr. Samudra Vijay, and the staff at ARCADIS helped with development of this documentation under EPA Contract
EP-C-09-027. Research Triangle Institute assisted with the development of this documentation under EPA
Contract EP-C-09-029.
Abstr	act
The United States Environmental Protection Agency (USEPA) has developed a model for the pulp and paper
sector that provides an integrated approach for investigating, developing, and evaluating strategies for reducing
the emissions of interest. This model is referred to as the Universal Industrial Sectors Integrated Solutions model
for the Pulp and Paper sector and was recognized as an integrated modeling tool by the Clean Air Act Advisory
Committee in its recent recommendations to USEPA. The model also was recognized by Resources forthe Future
which is a nonprofit and nonpartisan organization headquartered in Washington, DC, that conducts independent
research on energy, environment and natural resources. With inputs from users, the model can identify
technology options to meet various emission reduction strategies, provide estimates of the costs of these
options, and indicate the potential economic responses that may be provided by the industry to accomplish
these strategies. This document includes an introduction from the pulp and paper sector, data collection,
mathematical modeling framework of the model, the objectives of the model, etc. Analysis examples are
included to demonstrate how design of the model and its implementation strategies can handle the complex
interactions between economic concerns and the environment successfully, thereby overcoming the techno-
economic and emission-reduction challenges of multi-product, multi-market, and multi-pollutant sector-based
analyses.
iii

-------
Table of Contents
List of Appendices	vii
List of Tables	viii
List of Acronyms and Chemical Symbols	xii
1.	Introduction	1-1
1.1.	Pulp and Paper Facility Classifications	1-3
1.1.1.	Overview of Manufacturing Processes	1-5
1.1.2.	Chemical Pulping	1-6
1.1.3.	Mechanical Pulping	1-7
1.1.4.	Semi-Chemical Pulping	1-7
1.1.5.	Paper Recycling	1-7
1.2.	Energy Use in the Pulp and Paper Industry	1-8
1.3.	Emissions from the US Pulp and Paper Industry	1-11
1.4.	Overview of Universal ISIS-PNP	1-13
1.4.1. Pulp and Paper Modeling in Universal ISIS	1-15
1.5.	References	1-16
2.	Emissions Sources	2-1
2.1.	Background	2-1
2.2.	Air Emissions Sources	2-2
2.3.	Boilers	2-5
2.3.1.	Boiler Design and Fuels	2-5
2.3.2.	Source of Boiler Emissions	2-8
2.3.2.1.	Boiler NOx Emissions	2-8
2.3.2.2.	Boiler SO2 Emissions	2-8
2.3.2.3.	Boiler PM Emissions	2-9
2.3.2.4.	Boiler GFIG Emissions	2-9
2.3.3.	Boiler Emission Reduction Strategies	2-10
2.3.3.1.	Boiler NOx Reduction	2-10
2.3.3.2.	Boiler SO2 Reduction	2-17
2.3.3.3.	Boiler PM Reduction	2-20
2.3.3.4.	Boiler CO2 Reduction	2-24
2.4.	Recovery Furnaces	2-26
iv

-------
2.4.1.	Recovery Furnace Design and Fuels	2-26
2.4.2.	Source of Recovery Furnace Emissions	2-27
2.4.2.1.	Recovery Furnace NOx Emissions	2-28
2.4.2.2.	Recovery Furnace SO2 Emissions	2-28
2.4.2.3.	Recovery Furnace PM Emissions	2-28
2.4.2.4.	Recovery Furnace GFIG Emissions	2-29
2.4.3.	Recovery Furnace Emission Reduction Strategies	2-29
2.4.3.1.	Recovery Furnace NOx Reduction	2-29
2.4.3.2.	Recovery Furnace SO2 Reduction	2-32
2.4.3.3.	Recovery Furnace PM Reduction	2-33
2.4.3.4.	Recovery Furnace CO2 Reduction	2-35
2.5.	Lime Kilns	2-36
2.5.1.	Design and Fuels	2-36
2.5.2.	Sources of Emissions	2-37
2.5.2.1.	Lime Kiln NOx Emissions	2-37
2.5.2.2.	Lime Kiln SO2 Emissions	2-38
2.5.2.3.	Lime Kiln PM Emissions	2-38
2.5.2.4.	Lime Kiln GFIG Emissions	2-38
2.5.3.	Lime Kiln Emission Reduction Strategies	2-39
2.5.3.1.	Lime Kiln NOx Reduction	2-39
2.5.3.2.	Lime Kiln SO2 Reduction	2-40
2.5.3.3.	Lime Kiln PM Reduction	2-41
2.5.3.4.	Lime Kiln CO2 Reduction	2-43
2.6.	References	2-45
3. Universal ISIS-PNP Modeling Framework	3-1
3.1.	Introduction	3-1
3.2.	Objective Function	3-2
3.3.	Production and Costs in Universal ISIS-PNP	3-4
3.4.	Modeling Framework Architecture	3-8
3.5.	Constraints and Limitations	3-10
3.6.	Optimization and Post-Processing	3-11
3.7.	References	3-12
v

-------
4.	Pulp and Paper Data	4-1
4.1.	Data Collection Methodology	4-2
4.2.	Finished Product Data	4-4
4.2.1.	Data Processing	4-4
4.2.2.	Data Summary	4-6
4.3.	Mill Data	4-7
4.3.1.	Data Processing	4-7
4.3.2.	Process Characterization	4-9
4.3.3.	Boiler and Fuel Characterization	4-13
4.3.4.	Data Summary	4-14
4.4.	Cost Data	4-18
4.4.1.	Raw Material	4-18
4.4.2.	Maintenance and Repair	4-18
4.4.3.	Labor	4-19
4.4.4.	Fuel	4-19
4.4.5.	Electricity	4-19
4.4.6.	Solid Waste Disposal	4-19
4.4.7.	Transportation and Interregional Trade	4-19
4.5.	Emissions and Controls Data	4-20
4.5.1.	NOx	4-20
4.5.2.	S02	4-22
4.5.3.	CC^and Energy Efficiency	4-22
4.6.	Import Modeling Data	4-24
4.7.	Scenario Parameters	4-25
4.8.	References	4-26
5.	Model Calibration	5-1
5.1.	Methodology	5-1
5.2.	Data Used	5-2
5.2.1.	Prices	5-2
5.2.2.	Production	5-2
5.2.3.	Demand	5-2
5.2.4.	Imports to the United States	5-3
vi

-------
5.3.	Results	5-3
5.4.	Recommendations	5-4
6.	Illustrative Analysis	6-1
6.1.	Fuel Substitution	6-1
6.2.	Installation of Controls	6-3
6.3.	Implementation of Energy Efficiency Measures	6-4
6.4.	Summary	6-5
6.5.	References	6-6
7.	Universal ISIS-PNP Manual	7-1
7.1.	Hardware/Software Requirements of Universal ISIS-PNP	7-1
7.2.	Installation of GAMS (Supporting Software)	7-1
7.3.	Opening and Running Universal ISIS-PNP	7-3
7.3.1.	Open Project	7-4
7.3.2.	Open and Run Database	7-4
7.3.3.	Open and Run Model	7-6
7.4.	Model Input Requirements	7-8
7.5.	Pre-Processing of Data	7-9
7.6.	Output Database	7-10
7.7.	Running a Scenario in Universal ISIS-PNP	7-11
7.7.1.	Business as Usual (BAU)	7-11
7.7.2.	Scenario I - Emission Constraints	7-12
7.7.3.	Scenario II- Fuel Constraints	7-13
7.8.	Troubleshooting	7-14
7.8.1.	Compilation Errors	7-14
7.8.2.	Execution Errors	7-15
List of Appendices
Appendix A RTI International Draft Memo for Pulp and Paper Industry January 16, 2009
Appendix B AndoverTechnology Partners Memo March 15, 2010
Appendix C RTI International Memos for Paper Machine November 16, 2011 and March 29, 2013
vii

-------
I ist of Tables
Table 1-1. General Classification of Pulping Processes	1-4
Table 1-2. Major Paper Manufacturing Processes (DOE, 2005)	1-9
Table 1-3. Fuel Use for Pulp and Paper Production in 2000 (DOE, 2005)	1-10
Table 1-4. General Classification of Pulping Processes (DOE, 2005)	1-11
Table 1-5. Nationwide SO2 and NOx Emissions from Pulp and Paper Mills (Pinkerton, 2007)	1-12
Table 1-6. Trends in Nationwide SO2 and NOx Emissions from Pulp and Paper Mills (AF&PA, 2009)	1-13
Table 2-1. 2010 Emissions, 103 tons (NCASI, 2012)	2-3
Table 2-2. Nationwide GHG Emissions from the Pulp and Paper Manufacturing Industry	2-3
Table 2-3.	Pulp and Paper Sector — GHG Emissions Reported to the GHG Reporting Program for 2012 .2-4
Table 2-4. Relative Nitrogen and Sulfur Content of Fuels (NCASI, 2009a)	2-8
Table 2-5. Boiler NOx Control Technologies (NCASI, 2009a)	2-10
Table 2-6. Applicability of NOx Control Technologies (Andover, 2010)	2-16
Table 2-7. Boiler SO2 Control Technologies (NCASI, 2009a)	2-18
Table 2-8. Boiler PM Emission Control Technologies	2-22
Table 2-9. Representative Boiler Efficiency and GHG Emission Factors	2-24
Table 2-10. Recovery Furnace NOx Control Technologies	2-31
Table 2-11. Recovery Furnace SO2 Control Technologies	2-33
Table 2-12. Recovery Furnace PM Control Technologies	2-35
Table 2-13. Lime Kiln NOx Control Technologies	2-40
Table 2-14. Lime Kiln SO2 Control Technologies	2-41
Table 2-15. Lime Kiln PM Control Technologies	2-43
Table 4-1. Boiler Fuel Efficiency	4-1
Table 4-2. US PNP Sector Energy Use in 2000	4-2
Table 4-3. Universal ISIS Product Categories for the US Pulp and Paper Market	4-5
Table 4-4. Summary of Facilities Producing Each Major Product Category	4-6
Table 4-5. Summary of Production by Major Product Category	4-7
Table 4-6. Universal ISIS-PNP and RISI Region Comparison	4-9
Table 4-7. Production Capacity Example	4-10
Table 4-8. Example Product Summary Table for an Integrated Facility	4-11
Table 4-9. Example Product Summary Table for a Non-Integrated Facility	4-12
Table 4-10. Recycled Fiber Recovery Rates for Major Product Grades	4-14
Table 4-11. Fuel Availability Summary	4-15
Table 4-12. Recovery Furnace Controls Summary	4-16
viii

-------
Table 4-13.	Lime Kiln Controls Summary	4-16
Table 4-14.	Boiler Controls Summary	4-17
Table 4-15.	Applicability of NOx Reduction Technologies	4-21
Table 4-16.	CO2 Production from Combustion of Various Fuels (Ib/MMBtu)	4-22
Table 4-17.	CO2 Emission Factors for Combustion Sources at Pulp and Paper Mills (USEPA, 2009)	4-23
Table 4-18.	Energy Efficiency Measures for Pulp and Paper Industry Boilers	4-24
Table 5-1.	Reported Annual Prices of Paper Products	5-2
Table 5-2.	Reported Paper Products Annual Production	5-2
Table 5-3.	Reported Annual Paper Products Demand	5-3
Table 5-4.	Reported Annual Import Quantities of Paper Products	5-3
Table 5-5.	Reported and Calculated Prices of Products for 2007	5-3
Table 5-6.	Reported and Calculated Prices of Products for 2008	5-4
Table 5-7.	Reported and Calculated Prices of Products for 2009	5-4
Table 6-1.	Emission Intensity of Fuels	6-1
Table 7-1.	System Requirements for Software Installation	7-1
ix

-------
I ist of Figur	es
Figure 1-1. Historical Pulp and Paper Trends for the US and China	1-3
Figure 1-2. General Process Scheme for the Pulp and Paper Production Industry	1-5
Figure 1-3. Schematic of the Integrated and Non-Integrated Processes	1-6
Figure 1-4. Integrated View of Pollution Generation Pathways, Emissions Abatement Approaches, and
Multimedia Impactsforan Industrial Sector	1-14
Figure 1-5. Universal ISIS-PNP Modeling Framework	1-15
Figure 2-1. Stationary Combustion and Pulp Production Sources of Non-Biogenic CO2 Emissions	2-5
Figure 2-2. Stationary Combustion and Pulp Production Sources of Total CO2 (biogenic and non-biogenic
C02)	2-5
Figure 2-3. Fuels Used by Boilers in Pulp and Paper Sector by Fleat Input (NCASI, 2012)	2-6
Figure 2-4. FGD-Only Costs among 49 FGD Systems (NCASI 2009a)	2-19
Figure 2-5. Simplified Representation of the Kraft Pulping and Chemical Recovery System (USEPA
,2010a)	2-26
Figure 3-1.	Total Surplus in a Market	3-1
Figure 3-2	Stepwise Integration of the Inverse Demand Curve	3-4
Figure 3-3.	Modular Architecture of Universal ISIS-PNP	3-9
Figure 3-4.	Input and Output Data Management in Universal ISIS-PNP	3-10
Figure 4-1.	Methodology of Product Aggregation into Product Categories	4-5
Figure 4-2.	Mill Capacity Regions in the US (USDA, 1994)	4-8
Figure 4-3.	Example Boiler Data	4-14
Figure 4-4.	Domestic Transport of Pulp from Pulp Mills to Paper Mills	4-20
Figure 4-5.	Import Network of Pulp from Canada and/or ROW to the US	4-25
Figure 6-1.	Comparison of Base Emissions with Projected Emissions after Fuel Substitution	6-2
Figure 6-2.	Fuel Cost in Base Case and Fuel Substitution Scenarios	6-3
USEPA (2010). US Environmental Protection Agency. Available and Emerging Technologies for Reducing
Greenhouse Gas Emissions from the Pulp and Paper Manufacturing Industry. Available at:
http://www.epa.gov/nsr/ghgdocs/pulpandpaper.pdf. Last accessed on October 31, 2014	6-6
Figure 7-1.	GAMS Setup Wizard	7-2
Figure 7-2.	Select Destination Location	7-2
Figure 7-3.	GAMS Setup	7-3
Figure 7-4.	GAMS Setup	7-3
Figure 7-5.	Universal ISIS-PNP Project Opens on GAMS	7-4
Figure 7-6.	Database GAMS File in Universal ISIS-PNP Folder	7-5
Figure 7-7.	GAMSTitle Window before Running Database	7-5
x

-------
Figure 7-8. Model Run Status	7-6
Figure 7-9. Choosing Universal ISIS-PNP GAMS File	7-6
Figure 7-12. Universal ISIS-PNP Output File Location	7-8
Figure 7-13. Universal ISIS-PNP Data Structure	7-10
Figure 7-14. Input and Output Database Files of Universal ISIS-PNP	7-11
Figure 7-17. Applying Policy Parameters in Input File	7-13
Figure 7-18. Target Emissions Constraint in Input File	7-13
Figure 7-19. Fuel Constraints in Input File for both Pulp and Paper Mills	7-14
Figure 7-20. Insufficient Memory Error in Universal ISIS-PNP Run	7-15
Figure 7-21. Infeasibility Error in Universal ISIS-PNP Run	7-16
xi

-------
List of Acronyms and Chemical Symbols
ADMT
Air-dried Metric Ton
AFBC
Atmospheric Fluidized Bed Combustion
BACT
Best Available Control Technology
BAU
Business as Usual
BDST
Bone Dry Short Tons
BLS
Black Liquor Solids
Btu
British Thermal Unit
BXJ
Boxboard and Other Board
CaC03
Calcium Carbonate (limestone)
CaO
Calcium Oxide (Quicklime)
ch4
Methane
CNT
Containerboard
CO
Carbon Monoxide
C02
Carbon Dioxide
C02e
C02 Equivalents
COR
Corrugating Medium
CPW
Coated Printing and Writing Paper
CTMP
Chemi-thermo-mechanical Pulping
CYC
Cyclone
DBESP
Dry-Bottom Electrostatic Precipitator
ESP
Electrostatic Precipitator
FGD
Flue Gas Desulfurization
FGR
Flue Gas Recirculation
FST
Finished Short Ton
GAMS
General Algebraic Modeling System
GHG
Greenhouse Gas
HAP
Hazardous Air Pollutant
HHV
Higher Heating Value
ICR
Information Collection Request
ISIS
Industrial Sectors Integrated Solutions
kWh
Kilowatt Hour
LNB
Low NOx Burner
LoTOx
Low-Temperature Oxidation
LWS
Lime/Limestone Wet Scrubbing
MACT
Maximum Achievable Control Technology
MM Btu
Million British Thermal Units
xii

-------
n2
Nitrogen
n2o
Nitrous Oxide
Na
Sodium
Na2C03
Sodium Carbonate
Na2S
Sodium Sulfide
Na2S04
Sodium Sulfate
NaCI
Sodium Chloride
NAICS
North American Industry Classification System
NaOH
Sodium Hydroxide
NCASI
National Council for Air and Stream Improvement
NCG
Non-Condensable Gas
NESHAP
National Emission Standards for Hazardous Air Pollutants
NO
Nitric Oxide
NOx
Nitrogen Oxides
NSPS
New Source Performance Standard
O&M
Operation and Maintenance
OFA
Overfire Air
PC
Pulverized Coal
PFD
Process Flow Diagram
PIP
Packaging and Industrial Paper
PM
Particulate Matter
PM2.5
Particles Less Than or Equal to 2.5 Microns
PMio
Particles Less Than or Equal to 10 Microns
PNP
Pulp and Paper
ppm
Part(s) Per Million
PSC
Pulp Supply Center
RISI
Resource Information Systems, Inc.
ROW
Rest of the World
RMP
Refiner Mechanical Pulp
RSCR
Regenerative Selective Catalytic Reduction
SC
Supply Center
SCBR
Scrubber
SCR
Selective Catalytic Reduction
SDA
Spray Dryer Absorber
SNCR
Selective Non-Catalytic Reduction
S02
Sulfur Dioxide
S03
Sulfur Trioxide
TDF
Tire-Derived Fuel

xiii

-------
TMCP
Thermo -mechanical -chemi Pulping
TMP
Thermo-mechanical Pulp
TRS
Total Reduced Sulfur
ISIS
Industrial Sectors and Integrated Solutions (model)
ULNB
Ultra Low-NOx Burner
UPW
Uncoated Printing and Writing Paper
USEPA
US Environmental Protection Agency
VOC
Volatile Organic Compound
xiv

-------
Conver	sion Table - English Units to SI Units
To Obtain
From
Multiply by
m
ft
0.3048
m2
ft2
9.29 x 10-2
m3
ft3
2.83 x 10-2
°C
°F
5/9 x (°F - 32)
kg
lb
0.454
J/kg
Btu/lb
1.33 x 10-4
m3/s
Cfm
4.72 x 10-4
m3/s
Gpm
6.31 x 10-5
J/kWh
Btu/kWh
1054.8
mills
$
0.001
kg/m2
inches of Hg
345.31
metric ton
short ton*
0.907
'Note: in this document, "ton" refers to short ton and equals 2000 lb or 907 kg, unless otherwise specified.
"Metric ton" equals 1000 kg.
XV

-------
1.	1
In the National Academy of Science's 2004 report, "Air Quality Management in the United States," the National
Research Council recommended to the US Environmental Protection Agency (USEPA) that standard setting,
planning, and control strategy development should be based on integrated assessments that consider multiple
pollutants, and that these integrated assessments should be conducted in a comprehensive and coordinated
manner (NAS, 2004). With these recommendations, USEPA began to transition to establishing multi-pollutant
and sector-based approaches to manage air quality and environmental protection. The benefits of multi-
pollutant and sector-based analyses include the ability to identify optimal strategies that consider feasibility,
costs, and benefits across all pollutant types such as criteria air pollutants, hazardous air pollutants (HAPs), and
greenhouse gases (GHGs), while streamlining administrative and compliance complexities and reducing
conflicting and redundant requirements.
The development of policy options for managing emissions and air quality can be made more effective and
efficient through sophisticated analyses of relevant technical and economic factors. Such analyses are greatly
enhanced by the use of an appropriate modeling framework and, as a result, the Universal Industrial Sectors
Integrated Solutions (Universal ISIS) model has been developed at USEPA (ARCADIS, 2010). The Universal ISIS
was first populated with US cement manufacturing data (Universal ISIS-Cement), with subsequent efforts aimed
at building a representation of the US pulp and paper sector (Universal ISIS Pulp and Paper [Universal ISIS-PNP]
model). This document describes the framework of USEPA's Universal ISIS-PNP developed in the General
Algebraic Modeling System (GAMS) and its application to the US pulp and paper industry.
The US pulp and paper industry is a diverse sector that utilizes a variety of pulping processes and manufactures
hundreds of different grades of paper (DOE, 2005). The industry is grouped under paper manufacturing (North
American Industry Classification System [NAICS] code 322) and includes pulp, paper, and paperboard mills
(NAICScode 3221) and converted paper product manufacturing (NAICS code 3222). Pulp, paper, and paperboard
mills are facilities primarily engaged in producing pulp and/or paper and paperboard; paperboard is
distinguished from paper as a thicker product (>0.3 mm) but is manufactured in a similar manner (USEPA, 2002).
A facility primarily engaged in producing pulp is considered a pulp mill (NAICS code 32211), whereas a facility
primarily engaged in converting pulp into paper or paperboard is a paper mill (NAICS code 32212) or paperboard
mill (NAICS code 32213). A facility producing pulp and making paper with paper as the primary product is
considered an integrated mill and is classified as a paper or paperboard mill. Converted paper product
manufacturing includes facilities primarily using paper and/or paperboard products as a raw material to produce
paper-derived products (e.g., cardboard) that are not typically engaged in pulping or papermaking.
The US Census estimated there were 561 pulp, paper, and paperboard mills (NAICS code 3221) in the 2002
Economic Census, 32 of which were classified as pulp mills. In the 2007 Economic Census, there were 514 mills,
39 of which were classified as pulp mills. In the early 1980s, 40 percent of paper mills and 33 percent of
paperboard mills were integrated with pulp mills. By 1992, these numbers had fallen slightly to 38 percent and
29 percent, respectively (USDOC, 1996). However, more recently the industry has begun to move toward
integrated mills (DOE, 2005). The database used for Universal ISIS-PNP contains 514 facilities currently in
operation.
In 2009, the United States was the world's leading producer, consumer, and exporter of pulp and paper products
(RTI, 2009). Domestic production of paper and paperboard was 78.3 million tons in 2009 with a projected 2010
1-1

-------
production of 82.6 million tons (USDA, 2011). The 2007 US economic census estimates that the pulp, paper, and
paperboard industry (code 3221) produced $80 billion in revenue, and that US production of pulp, paper and
paperboard experienced a sharp decline during 2008-2009 associated with the global financial crisis. Production
has recovered from these lows but has failed to reach the previous production peak of 2007. Capacity utilization
declined during the 2008-2009 crises but has recovered to nominally 84 percent in line with capacity utilization
in 2007. Recovery of capacity utilization with lower overall production suggests shutdowns and capacity
reductions in the industry (UNECE, 2011).
Exports of paper and paperboard were 43.9 million tons in 2009 while imports were 20.3 million tons (USDA,
2011). US paper and paperboard exports exceeded imports in 2009, and the US remained a net exporter through
the first half of 2011 (UNECE, 2011). Canada leads in shipping newsprint to this country while the United States
predominates in wood pulp exports to Canada (MFI, 1998), and Canada is the industry's largest trading partner;
21.9 million tons of pulp, paper, and paperboard flowed between the two countries in 2001. Exports of pulp and
paper products to China, Japan, Europe, South America and Mexico have been increasing steadily. Exports of
pulp to China, Japan and Korea were valued at more than $700 million in 2004 (DOC, 2004).
Historical production, export, and import data were obtained from the Food and Agriculture Organization of the
United Nations for "pulp for paper" and "paper and paperboard." In 2011, the five biggest world producers of
pulp for paper were the US (50.2 million metric tons), China (21.1 million metric tons), Canada (18.3 million
metric tons), Brazil (13.9 million metric tons), and Sweden (11.7 million metric tons). The largest importers of
pulp were China (14.0 million metric tons), the US (5.3 million metric tons), Germany (4.6 million metric tons),
Italy (3.5 million metric tons), and Korea (2.5 million metric tons). The largest exporters of pulp were Canada
(9.2 million metric tons), Brazil (8.5 million metric tons), the US (8.3 million metric tons), Chile (4.0 million metric
tons), and Indonesia (2.9 million metrictons). In 2011, the five highest world producers of paperand paperboard
were China (103.1 million metric tons), the US (77.4 million metric tons), Japan (26.2 million metric tons),
Germany (22.7 million metric tons), and Canada (12.1 million metric tons). The largest importers were Germany
(10.5 million metric tons), the US (9.4 million metric tons), the United Kingdom (6.9 million metric tons), France
(5.6 million metric tons), and China (5.2 million metric tons), and the largest exporters were the US (13.9 million
metric tons), Germany (13.3 million metric tons), Finland (10.5 million metric tons), Sweden (10.5 million metric
tons), and Canada (9.1 million metric tons). Historically, the US and China have been the leaders in pulp and
paper production. China production of paper began to increase significantly in 2002 and surpassed US
production in 2008, as shown in Figure 1-la. Pulp production did not follow this trend, as the US maintains a
significantly higher pulp production for all of the years analyzed (1980-2011), as shown in Figure 1-lb. As
expected with China's large increase in paper production but minimal increase in pulp production, China's pulp
imports increased significantly from 2002 to 2011 (Figure 1-lc). The US's exports of pulp also increased during
this time (Figure 1-ld).
1-2

-------
1.2E+8
& 1.0E+8
u
-------
form called market pulp from an integrated mill or from a pulp mill or purchases recycled paper. A non-
integrated pulp mill produces pulp and sells it to otherfacilities for conversion into paper. Integrated mills share
common systems for generating steam and energy and for treating wastewater and eliminate transportation
costs for acquiring pulp. Non-integrated mills must obtain pulp from another source but are typically smaller
and can be located in urban locations (MGH 1999; AF&PA 1998; Paperloop, 2003; Saltman, 1998). Figure 1-2
shows a general process scheme for the pulp and paper production industry.
Table 1-1. General Classification of Pulping Processes
Category
Chemical
Semi-Chemical
Mechanical
Description
Pulping with chemicals and heat
(little or no mechanical energy)
Pulping with combinations of
chemical and mechanical
treatments
Pulping by mechanical energy (small
amount of chemicals and heat)
Yield a
Lower yield (45-50 % for bleachable
or bleached pulp, 65-70 % for brown
papers)
Intermediate yield (55-85 %)
High yield (85-96 %) (lignin not
removed)
Wood Used
All woods (kraft); some hardwoods
and non-resinous softwoods (sulfite)
Mostly hardwoods
Non-resinous softwoods, some
hardwood like poplar
Pulp
Properties
High strength
High water absorption
Low brightness
"Intermediate" pulp properties
Good stiffness and moldability
Low strength
High brightness
High opacity, softness, and bulk
Good print quality
Major
Processes
Kraft (sulfate)
Sulfite
Neutral sulfite semi-chemical
High-yield kraft
High-yield sulfite
Stone ground-wood
Refiner mechanical pulp
Thermo-mechanical pulp
Chemi-thermo-mechanical pulp
Products
Kraft: bag, wrapping, linerboard,
bleached pulps for white writing and
printing papers
Sulfite: fine paper, tissue, glassine,
newsprint, dissolving pulp
Corrugating medium
Food packaging board
Newsprint, magazine
Newsprint, magazines, catalogs
Books
Container board
a. Yield = weight of pulp produced (oven dry) divided by weight of original wood (oven dry).
1-4
-------
Air Emissions & Waste



i
1
i



\
Pulping
Process
Pulp ^
Paper
Production
1
Products
i

1
A


A
I
-Non-Integrated Paper I
-Integrated Pulp & Paper I
Figure 1-2. General Process Scheme for the Pulp and Paper Production Industry
1.1.1. Overview of Manufacturing Processes
Integrated pulp mills produce paper using six general processing stages: wood preparation, cooking or pulping,
pulp washing, pulp screening, bleaching (optional, depending on product), and paper making.
Wood preparation: The wood preparation process involves wood cutting, transporting, debarking, chipping, and
screening of the wood material. Hardwoods and softwoods can be harvested from tree plantations or from
forests and species vary based on harvesting location. Wood is delivered to the pulp mill in one of two ways:
logs or sawmill chips (residuals from sawmills). The logs are trimmed to appropriate processing lengths and the
bark is removed (i.e., debarking) and burned in hog fuel boilers or sold for landscaping purposes. After debarking,
the logs are reduced to chips that are the appropriate size for pulping. The chips are screened to remove
oversized chips and sawdust. Oversized chips are re-chipped until they are the appropriate size, and sawdust is
typically burned in a hog fuel boiler with the bark. From wood preparation, the chips proceed to pulping.
Cooking or pulping: The cooking (pulping) process is where the wood is broken down into fibers that can be used
for papermaking. In the case of chemical pulping, cellulosic fibers are separated from the cellulose-
hemicellulose-lignin matrix in wood using high temperatures, pressure, and chemicals. In the case of mechanical
pulping, logs are chipped and mechanically broken into smaller pieces. These pulping methods are discussed in
more detail in the following section. From the pulping process, pulp and spent cooking liquor proceed to pulp
washing.
Pulp washing: The pulp washing process is used to remove cooking chemicals and the dissolved wood
components in the cooking liquor for recovery and for energy generation. The recovery of these materials may
also minimize the addition of chemicals and solids to the effluent treatment plant. The chemicals (inorganic and
organic) are separated (washed) from the cooked pulp and screened. Pulping and pulp washing steps are very
similar in kraft and sulfite processes. The pulp proceeds to pulp screening and the liquor proceeds to the
chemical recovery process.
Pulp screening: The pulp screening process separates cooked pulp fibers from uncooked fiber bundles and knots.
In the screening process, unwanted particles are removed by passing the pulp over pulp screens equipped with
fine holes or slots. These screens may operate using gravity, vibrations, centrifugal force, or pressure. The pulp
1-5
-------
proceeds from screening to bleaching if the final product requires bleached pulp or to the paperrnaking process
if the final product utilizes unbleached pulp.
Bleaching: The bleaching process involves removing the lignin that still remains after cooking (chemically
whitening) or breaking double bonds in the lignin without removing it (brightening), as the lignin contains the
chromophoric groups that make the pulp dark. Bleaching and cooking are both delignification processes, and
modern developments have tended to blur the difference between the two processes. However, traditionally
the term 'bleaching' is reserved for delignification that is taking place downstream of the cooking process. Not
all products require bleaching or the same amount of bleaching. Bleached pulp proceeds to the paperrnaking
process.
Paper making: The paper-making process involves stock preparation, dewatering, pressing, drying and finishing.
Pulp fibers are treated mechanically by refining to produce flexible fibers suitable for paperrnaking. These fibers
are blended with product specific additives (e.g., fillers for printing paper, or wet strength agents for tissue) and
are diluted significantly with water (<1 % fibers). This slurry is processed on a paper machine, which creates a
fiber mat and removes water by gravity, suction, pressure, and heat. The paper can be converted to final
products onsite, or may be shipped to another location for conversion.
These six processes could occur at an integrated facility (integrated pulp mill and paper mill). Alternatively, a
selection of them could occur at a stand-alone paper mill or a stand-alone pulp mill (non-integrated facility). For
example, a stand-alone paper mill could import pulp for the paperrnaking process (non-integrated paper mill),
as illustrated in Figure 1-3.
Cooking (Pulping)
Wood Preparation

Pulp Screening
Pulp Washing
Pulp mills
f
Bleaching
\
\

/
Paper Mills
Paperrnaking
t
Figure 1-3. Schematic of the Integrated arid Non-Integrated Processes
1.1.2. Chemical Pulping
In chemical pulping, wood chips are mixed with a chemical solution, heated under pressure to increase the
reaction rate, and then disintegrated into fibers. The chemical recovery process involves evaporation,
combustion, causticizing, and calcining. These processes are used to generate energy and recover cooking
1-6
-------
chemicals. The weak black liquor from pulp washing is processed using a multiple-effect evaporator system to
increase the black liquor solids content by removing water. This processing is done to improve the heating value
of the liquor, because it will be burned in a recovery furnace to generate steam. The purpose of the recovery
furnace is to burn the organics in the black liquor and recover the inorganics in molten form. These inorganics
(known as smelt) are dissolved to create green liquor. Green liquor is then clarified and causticized using lime to
create white liquor for the pulping process. Lime mud is collected from the white liquor clarifier and burned in
a lime kiln to regenerate lime for the caustization process.
Kraft chemical recovery pulping is by far the most common pulping process used by plants in the US for virgin
fiber and produced approximately 83 percent of all US virgin pulp tonnage during 2000 (USEPA, 2002). The kraft
pulping process uses alkaline cooking liquor (called white liquor) of sodium hydroxide (NaOH) and sodium sulfide
(Na2S) to digest wood, while the similar soda process uses only NaOH to digest the wood. The cooking liquor in
the sulfite pulping process is an acidic mixture of sulfurous acid (H2SO3) and salts of bisulfite ion (HSO3 ). The
counter ion used in sulfite cooking liquor preparation is typically calcium, although historical counter ions also
included ammonium, magnesium, and sodium.
1.1.3.	Me<. ¦ - v . :.
Mechanical pulping is the oldest methodology used to separate pulp fibers from the wood matrix. The stone
ground wood process was the most widely used mechanical pulping process until the 1990s. This method
produces pulp by pressing a log against a rotating stone at atmospheric pressure. Fibers and fiber fragments are
collected by washing the stone and are then processed. Pressurized ground wood, a similar process, uses the
same technology, but grinds the logs at a temperature higher than 100 °C.
Mechanical pulping technology eventually shifted towards RMP, and by 1990, half of the mechanical pulp in the
US was produced by this method. The advantage of RMP is that it uses wood chips instead of logs and refiner
plates instead of stones. Three additional processes, thermo refiner mechanical pulp, pressure refiner
mechanical pulp, and chemi-refiner mechanical pulp add pre-steaming of chips, increased refiner temperature,
and chemical treatment, respectively. Finally, thermo-mechanical pulping (TMP) modified the RMP process by
steaming chips under pressure prior to and during refining. Several variations of this process are utilized today,
including pressure/pressure thermo-mechanical pulping, chemi-thermo-mechanical pulping (CTMP), thermo-
mechanical-chemi pulping (TMCP), and long fiber chemi-mechanical pulping.
1.1.4.	5 . 1 ' 1 'I ' ' ' , v,
Semi-chemical pulping utilizes both chemical and mechanical defibrillation methods. Examples of semi-chemical
processes are neutral sulfite semi-chemical, high-yield kraft, and high-yield sulfite. The high-yield chemical
processes utilize minimal kraft and sulfite chemical cooking followed by mechanical defibrillation. The neutral
sulfite semi-chemical process is the most widely utilized semi-chemical process and is typically used to process
hardwood. The liquorfrom these processes can be recovered in the kraft recovery furnace if they are collocated
at a facility that uses kraft pulping, or a fluidized bed incinerator can be used.
1.1.5.
In the fiber recycling process, pulp fiber is recovered from previously manufactured products such as cardboard
or office paper. There are five basic grades of wastepaper that are commonly collected: mixed paper, old
newsprint, old corrugated container, pulp substitutes, and high-grade de-inked. Mixed paper is the category that
includes office waste, boxboard cuttings, and other grades. Pulp substitutes include unprinted and uncoated
1-7
-------
paper and board. High-grade de-inked is printing and writing grades that have previously been printed (IPST,
2006).
These waste papers must be collected and transported to a processing facility, which can be expensive. Typically,
plants that utilize wastepaper are located in urban areas where an abundant supply is available. After collection,
waste papers are re-pulped using water and agitation. Contaminants are removed from the pulp through
screening, de-inking, washing, and bleaching.
1.2. Energy Use in the Pulp and Paper Industry
Pulp and paper production is an energy intensive process. In 2002, the paper manufacturing industry consumed
over 2.4 quads (quadrillion or 1015 British thermal units [Btu]) of energy according to the Manufacturing Energy
Consumption Survey, and represented over 15 percent of US manufacturing energy use (MECS, 2003; DOE,
2005). Large electricity losses are incurred at offsite utilities during generation and transmission of electricity; if
these losses are included, the total energy associated with paper manufacturing reaches 2.8 quads (based on
conversion factor of 10,500 Btu/kilowatt hour [kWh]).
Fuels comprise the bulk of the industry's primary energy use with only 7 percent of the energy use being
purchased electricity. Nearly 55 percent of the energy demand is met by the use of biomass-based waste and
byproduct fuels (e.g., wood, spent pulping liquors, chips, sawdust, and bark). Despite its large use of biomass-
based fuels, the paper manufacturing industry is the fourth largest consumer of fossil energy, after chemicals,
petroleum refining and steel. Energy intensity of various stages of production is given in Table 1-2. Process
energy consumption can vary widely due to different technologies or variations in operating practices and
feedstock composition. Energy demand among pulping processes can be quite different.
The industry relies on a diverse fuel mix. To supplement the use of fossil fuels, the industry self-generates
electricity and heat using byproduct fuels such as wood, spent pulping liquors, chips, sawdust, and bark. In 2002,
over 50 percent of the industry's energy demand was self-generated through the use of biomass-based fuels.
The pulp and paper sector generates more electricity than any other manufacturing industry (51,208 kWh in
2002) (DOE, 2005).
Power boilers are often capable of being fired with multiple fuels. The design of power boilers varies with fuel
type (e.g., oil, gas, coal, bark). Some are designed to process the so-called "hog fuel," a mixture of wood material
generated onsite (e.g., bark, wood chips) that is constantly changing and is mill-dependent. Hog fuel boilers may
be supplemented with oil, coal, or natural gas (e.g., if fuel moisture is too high, or during disturbances in solid
fuel feeding). Non-integrated paper mills typically rely on fossil fuels because they do not produce wood
byproducts.
Table 1-3 presents the various purchased and self-generated fuels used by the industry.
1-8
-------
Table 1-2. Major Paper Manufacturing Processes (DOE, 2005)
Operation
Major Processes
Average energy (106
Btu/ton pulp)
Average energy
(106 Btu/ton
paper/paperboard)c
Wood Preparation
Debarking
Chipping and conveying
0.10
0.35
n/a

Chemical pulping
2.68


Kraft process
2.60

Pulping
Sulfite process
Semi-chemical pulping
Mechanical pulping
Recycled paper re-pulping
5.38
3.86
7.68b
1.30
n/a

Evaporation
3.86

Kraft Chemical
Recovery furnace
1.13a
n/a
Recovery
Re-causticizing
Lime kiln (calcining)
1.02
2.03
Bleaching
Mechanical or chemical pulp bleaching
2.3
n/a

Paper refining and screening

0.84

Forming, pressing, finishing and drying of:



Newsprint

5.61
Paper Making
Tissue
Uncoated paper
Coated paper
Linerboard
n/a
9.77
6.90
7.10
4.97
a.	Does not reflect energy generated by the recovery furnace, which ranges from 4-20 million Btu/ton pulp.
b.	Value for chemi-thermo-mechanical pulping.
c.	Includes energy from steam and electricity for each product except tissue, which includes steam, electricity, and fuel.
n/a=not available in the report
1-9
-------
Table 1-3. Fuel Use for Pulp and Paper Production in 2000 (DOE, 2005)
Fuel Source
Billion Btu Consumed
Percent of Total
Electricity
155,319.80
7
Steam
33,882.90
1.5
Coal
265,800.00
12
Petroleum Products
102,184.20
4.6
Natural Gas
395,611.00
17.7
Other
24,052.60
1.1
Excess Energy Sold
44,836.00

Total Purchased
932,014.50
43.9
SELF-GENERATED
Hog Fuel
327,359.00
14.7
Spent Liquor (solids)
894,985.90
40.3
Hydroelectric Power
4,989.70
0.2
Other
19,866.50
0.9
Total Self-Generated
1,247,201.10
56.1
Typically, a combustion unit (i.e., recovery furnace) is used to recover the cooking chemicals from spent cooking
solutions (or liquors). Although the primary purpose of the recovery furnace (sometimes referred to as a
recovery boiler) is to recover chemicals from spent pulping liquors (e.g., black liquor) for reuse, the recovery
furnace also produces heat used to generate steam and electricity. Recovery furnaces at kraft pulp mills burn
black liquor which has been concentrated through a multiple effect evaporator train and a direct contact or non-
direct contact evaporator prior to being fired. Kraft and soda mills have an additional chemical recovery process
in which a lime kiln is used to regenerate a portion of the chemical cooking solution.
Researchers are currently demonstrating gasification technologies that convert biomass and black liquor into a
synthesis gas (syngas), which can be combusted in a gas turbine to generate electricity. In combined-cycle
gasification, the gas turbine exhaust is then used to produce steam for generation of additional electricity or
process heat (DOE, 2005). Currently, black liquor gasification technologies are in operation at three US pulp mills
(two kraft mills and one stand-alone semi-chemical mill). Once black liquor gasification has been successfully
introduced, adoption of biomass gasification will likely follow.
In addition, the forest product industry is still hopeful that technologies for conversion of biomass to biofuels,
including gasification and hemicellulose conversion to ethanol, will continue to expand and will be able to extract
more energy from the same amount of biomass and thereby reduce the use of fossil fuels and their emissions.
Similarly, research continues on the production of renewable fuels at mills that could be used onsite to replace
natural gas in equipment such as lime kilns. Widespread deployment is dependent upon many factors,
particularly Federal research programs, the availability of capital, and successful scale-up from pilot operations
to commercial facilities. A new technology for black liquor combustion in a dual-pressure recovery boiler
promises significant improvement in steam generation and cogenerated electric power, which would reduce
1-10
-------
fossil fuel demands at mills and utilities and the associated emissions. The dual pressure boiler technology is
developed and waiting for full-scale commercial demonstration (AF&PA, 2009).
Table 1-4 provides a summary of the various pulp and papermaking processes and their relative energy
intensities (energy consumed per ton of pulp).
Table 1-4. General Classification of Pulping Processes (DOE, 2005)
Pulping
Process
Wood Pulp
Production
for 2001 (%)
Major Processes
Products
Average Energy
Intensity*
(106 Btu/ton pulp)
Chemical
54
Kraft (sulfate)
Bags, wrapping paper, linerboard,
newsprint, bleached pulp for white
writing and printing papers
Electricity: 0.50
Steam: 2.10
Total: 2.60
Sulfite
Fine paper, tissue, glassine, newsprint,
dissolving pulp
Total: 5.38
Semi-chemical
4
Neutral sulfite semi- chemical
High yield kraft
High yield sulfite
Corrugated board, food packaging
board, newsprint, magazine
Electricity: 1.56
Steam: 2.30
Total: 3.86
Mechanical
5
Stone ground wood
Refiner mechanical pulp
Thermo-mechanical pulp
Chemi-thermo-mechanical
pulp
Newsprint, magazine, catalogs, books,
container board
Electricity: 6.08
Steam: 1.60
Total: 7.68
Recycled
37
N/A
Newsprint, printing/writing paper,
tissue, packaging, containerboard,
paperboard
Electricity: 0.50
Steam: 0.80
Total: 1.30
* Electricity conversion factor of 3412 Btu/kWh.
1.3. Emissions from the US Pulp and Paper Industry
The environmental impacts from the pulp and paper industry can potentially come from hazardous chemicals,
thermal loading to natural waterways, odor, combustion, and solid wastes. The industry is in the process of
minimizing environmental impacts by increasing the use of recycled paper, improving energy efficiency, and
making capital investments for effective compliance with regulations.
The pulp and paper industry generates more than 12 million tons per year of solid waste, consisting primarily of
de-watered sludges. The standard treatment for these wastes in the past was to deposit them in landfills. Today
they are more often being handled by incineration, conversion to useful products, and land application. Most
solid waste from mills, such as sludge from de-inking plants, is non-hazardous and requires no special handling
(Paperloop, 2003).
A survey study estimated that boilers are the dominant emission source, accounting for nearly 90 percent of the
sulfur dioxide (SO2) and 70 percent of the nitrogen oxides (NOx) (NCASI, 2004).
Table 1-5 presents the 2005 SO2 and NOx emissions results of the National Council for Air and Stream
Improvement (NCASI) study (in tons per year). Kraft mill sources (primarily recovery furnaces) account for most
1-11
-------
of the remaining nationwide SO2 and NOx emissions from pulp and paper mills. Table 1-6 illustrates that the
emissions from sulfite and semi-chemical pulping operations are minimal compared to the same from kraft mills.
Based on a survey of pulp and paper mills conducted by NCASI, there were approximately 425 pulp and paper
mills that operated stationary combustion units (e.g., power boilers, recovery furnaces) in 2005 (NCASI, 2006).
All of these 425 mills fall under NAICS code 3221. Of these 425 mills, 129 produced chemical pulp (including 108
integrated kraft /soda pulp mills, 8 sulfite pulp mills, and 13 stand-alone semi-chemical pulp mills) and 19 were
mechanical pulp mills. The remainder of the mills operated combustion sources (e.g., power boilers) but did not
produce pulp (Pinkerton, 2007).
Table 1-5. Nationwide SO2 and NOx Emissions from Pulp and Paper Mills (Pinkerton, 2007)
Source
S02
(in thousands of
tons per year)
NOx
(in thousands of
tons per year)
Boilers
293
153
GasTurbines
-
3
Kraft Recovery Furnaces*
40
59
Kraft Smelt Dissolving Tanks*
1
1
Kraft Lime Kilns*
2
9
Kraft Thermal Oxidizers
2
1
Sulfite Pulp Mills
2
3
Semi-Chemical Pulp Mills
<1
1
TOTAL
340
230
¦"Includes units at one soda pulp mill.
1-12
-------
Table 1-6. Trends in Nationwide SO2 and NOx Emissions from Pulp and Paper Mills (AF&PA, 2009)

1980
1985
1990
1995
2000
2005
SO2 (in thousands of tons per year)
Boilers
730
523
461
393
351
293
Kraft pulping
122
153
96
86
57
44
Sulfite/Semi-
chemical pulping
23
23
14
8
4
3
TOTALS02
875
699
571
487
412
340
NOx (in thousands of tons per year)
Boilers
207
231
231
233
199
156
Kraft pulping
66
73
69
76
76
70
Sulfite/Semi-
chemical pulping
2
2
7
7
3
4
TOTAL NOx
275
306
307
316
278
230
1.4. Overview of Universal ISIS-PNP
The Universal ISIS-PNP, a sector-based linear programming model, is designed to facilitate the analyses of
emission reduction strategies for multiple pollutants while accounting for plant-level economic and technical
factors such as the type of emission units (for pulp and paper - power boilers, hog fuel boilers, recovery furnaces
and lime kilns), associated capacities, locations, costs of production, and applicable controls and costs. For each
of the emission reduction strategies under consideration, the Universal ISIS-PNP is able to identify optimal (least
cost) industry operation by selecting cost-effective controls to meet the demand for pulp and paper while
complying with emission reduction requirements over the time period of interest.
The design of Universal ISIS-PNP allows for incorporating multiple industries within a multi-market, multi-
product, multi-pollutant, and multi-region emissions trading framework. The objective function in Universal ISIS-
PNP maximizes total (consumer and producer) surplus and uses an elastic formulation of the demand function
to estimate area under the demand curve. The total surplus represents the difference between the cumulative
amount that consumers value a product and the cumulative costs of producing the product. Total surplus is
calculated for both Business as Usual (BAU) and policy cases. The change in total surplus between BAU and
specific policy cases may be used to evaluate societal costs of policy implementation against societal benefits
that may not be incorporated in the model. Emission reduction strategies are incorporated into the model
through various constraints depending on the type of strategy.
The Universal ISIS code is written in GAMS language. Input data from Universal ISIS-PNP, organized in various
spreadsheets of a Microsoft Excel workbook, are passed onto GAMS. These input data consist of an industry
database, which provides unit-level production, capacity, production cost, and emissions information. A controls
database provides information regarding applicable air pollution control technologies and their cost and
emission control characteristics. A policy module is used to specify various parameters of interest to the policy
analyst such as emissions cap, emission reduction scenarios, and discount rate. The input data, control data, and
policy parameters are then transmitted to the optimization components of the Universal ISIS, where they are
used to solve the selected baseline and policy cases. The results are post-processed to calculate values of various
1-13
-------
outputs of interest. The output data are exported to Excel spreadsheets for further analyses and graphical
representation of selected results.
Within an industrial sector, generally emissions arise from four pathways: (1) on-site emissions due to
combustion of fossil fuels for energy at plants, (2) on-site emissions due to processing of certain raw materials
(3) off-site emissions due to combustion of fossil fuels at power plants to generate the electricity needed by the
industrial sector, and (4) overseas emissions associated with imports. These pathways are depicted in Figure 1-
4.
Raw Materials
Pollution Reduction Policy
(Rate-based, Cap-and-trade,
Emissions Taxes) \
Pollution Controls
(NOx, C02, S02, VOC,
(Raw Material Substitution)
Water
Air Emissions
Industrial Sector
(Cement, Pulp & Paper, and
Iron & Steel etc.)
Fuel
(Fuel Substitution
Reduction)
Improvements
Waste
Electricity
(Change in FuelPower
Plant Emissions)
Products
(Production Substitution)
Imports
(Oversee Emissions)
Markets
Figure 1-4. Integrated View of Pollution Generation Pathways, Emissions Abatement Approaches, and
Multimedia Impacts for an Industrial Sector
Also shown in Figure 1-4 are the potential options for abating emissions from industrial sectors and multimedia
impacts. The options shown in green are pollution prevention measures, and the ones in red are mitigation
measures. Clearly, the integrated picture presented in Figure 1-4 makes a compelling case for considering
commodity production/supply activities along with emissions while developing holistic emission reduction
strategies. While developing the Universal ISIS-PNP framework, care has been taken to build the emission
pathways and abatement options shown in Figure 1-4. Example emission reduction policies that can be
evaluated using Universal ISIS-PNP are:
•	Criteria pollutants (NOx, SO2, particulate matter, carbon monoxide [CO]) -emission limits and/or cap-and-
trade
•	Flazardous Air Pollutants (e.g., total FIAPs, benzene, hydrogen chloride) - emission limits
1-14
-------
•	Carbon dioxide (CO2) - cap-and-trade and/or emission taxes
•	Long and short time horizons: CO2 (decades), criteria pollutants (annual)
Policies may be simulated over long and short time horizons such as a CO2 policy that occurs over a decadal
time-frame and a criteria pollutant policy that occurs on an annual-time frame. The Universal ISIS model is also
capable of evaluating requirements at a regional or national scale.
1.4.1. Pulp and Paper Modeling in Universal ISIS
The Universal ISIS-PNP modeling efforts for the pulp and paper sector are focused on the power boilers
(including hog fuel boilers), recovery furnaces, and lime kilns at integrated and non-integrated mills. The industry
database is comprised of 514 facilities, both integrated and non-integrated, populating US production capacity
in 2007. Both the emissions information and the controls database focus on HAPs, criteria air pollutants, and
greenhouse gases. Both databases can be updated as additional data are acquired and incorporated. An
overview of the Universal ISIS-PNP framework for the pulp and paper industry is presented in Figure 1-5.
Emissions related to pulping at non-integrated paper mills are incorporated as off-site emissions.
Onsite
emissions
emissions
PULP MILLS
PAPFR Ml! LS
.Transport
Material
Products
Recycling
Electricity
Boilers
Rest of the
World
Pulp Imports
and Exports
Rest of the
World
Imports and
Exports
Consumption
(Demand Centers)
Paper Mills
(Integrated or Non-
Integrated)
Pulp Mills
(Integrated or Non-
Integrated)
Figure 1-5. Universal ISIS-PNP Modeling Framework
1-15
-------
1.5. References
ARCADIS (2010). Development of USEPA Industrial Sector Integrated Solutions (ISIS) Model, Quality Assurance
Project Plan. Prepared under contract number EP-C-09-027 for the US Environmental Protection Agency, Air
Pollution Prevention and Control Division, Research Triangle Park, NC. February 3, 2010.
AF&PA (2009). Email communication from Tim Hunt, AF&PA to Beth Palma, USEPA, and transmitting AF&PA
responses to a list of USEPA requests from the Pulp and Paper industry for multi-pollutant/economic modeling.
May 22, 2009.
AF&PA (1998). American Forest & Paper Association, Paper, Paperboard & Wood Pulp: 1998 Statistics Data
through 1997. 1998.
DOC (2004). US Department of Commerce, US Census Bureau, Foreign Trade Statistics. November 2004.
http://www.census.gov/foreign-trade/statistics/highlights/top/top0411.html. Last accessed on October 31,
2014.
DOE (2005). Department of Energy. Energy and Environmental Profile of the US Pulp and Paper Industry. US DOE
Office of Energy Efficiency and Renewable Energy (EERE), Industrial Technologies Program. December 2005.
IPST (2006). Institute of Paper Science and Technology (IPST) at Georgia Institute of Technology, Recycling In The
Paper Industry. 2006.
MECS (2003) Manufacturing Energy Consumption Survey. 2003. http://www.eia.doe.gov/emeu/mecs/. Last
accessed on October 9, 2014.
MFI (1998). Miller Freeman, Inc., 1999 North American Pulp & Paper Fact Book. 1999.
MGFI (1999). US Department of Commerce and International Trade Administration, US Industry & Trade
Outlook®'99. 1999.
NAS (2004). Air Quality Management in the United States. National Research Council (US), Committee on Air
Quality Management in the United States, National Academies Press, Washington, 2004.
http://books.nap.edu/catalog.php7record id=10728, accessed October 21, 2008.
NCASI (2004). National Council for Air and Stream Improvement. Compilation of Criteria Air Pollutant Emissions
Data for Sources at Pulp and Paper Mills Including Boilers. Technical Bulletin No. 884. August 2004.
NCASI (2006). National Council of the Paper Industry for Air and Stream Improvement, Inc. Pulp and Paper Mill
Emissions of SO2, NOx, and Particulate Matter in 2005. NCASI Special Report No. 06-07. December 2006.
Paperloop (2003). Paperloop, Pulp & Paper 2002 North American Fact Book, 2003.
Pinkerton (2007). J.E.. Sulfur Dioxide and Nitrogen Oxides Emissions from US Pulp and Paper Mills, 1980-2005.
Journal of the Air & Waste Management Association. 57(8):901-906, August 2007.
RTI (2009), Memorandum from Katie Flanks, RTI, to Beth Palma, USEPA, January 16, 2009.
1-16
-------
Saltman, D., L.M. Thompson, and K.M. Bennett (1998), Pulp and Paper Primer 2nd Edition. TAPPI Press, Atlanta
GA, 1998.
UNECE (2011). Forest Products Annual Market Review 2010 - 2011. United Nations Economic Commission for
Europe/Food and Agricultural Organization of the United Nations, Geneva, 2011.
USDA (2011) US Department of Agriculture (2011): Annual production and demand data for uncoated free sheet,
corrugating medium and solid bleached board products, March, 2011.
http://www.usda.gov/wps/portal/usda/usdahome?navid=DATA STATISTICS&navtype=RT&parentnav=PRODUC
ERS Last accessed November 11, 2014.
USDOC (1996). US Department of Commerce, 1996 Annual Survey of Manufacturers: Statistics for Industry
Groups and Industries. 1996, M96(AS)-1. http://www.census.gOv/prod/3/98pubs/m96-asl.pdf . Last accessed
on November 11, 2014.
USEPA (2002). Profile of the Pulp and Paper Industry - 2nd edition, Office of Compliance, Office of Enforcement
and Compliance Assurance, U.S. Environmental Protection Agency, Washington, DC, EPA/310-R-02-002,
November 2002.
USEPA (2010). US Environmental Protection Agency. Office of Air and Radiation. Available and Emerging
Technologies for Reducing Greenhouse Gas Emissions from the Pulp and Paper Manufacturing Industry.
Available at: http://www.epa.gov/nsr/ghgdocs/pulpandpaper.pdf. Last accessed on October 31, 2010.
1-17
-------
2. Emissions Sour	ces
2.1. Background
Pulp and paper manufacturing processes emit a variety of air pollutants that are regulated by federal air emission
standards and permitting limitations. Emissions of total reduced sulfur (TRS), malodorous compounds
characteristic of kraft pulp mills, are regulated underfederal new source performance standards (NSPS) for kraft
pulp mills and state limitations based on federal emission guidelines for kraft pulp mills. Emissions of particulate
matter (PM) from kraft chemical recovery combustion sources - recovery furnaces, lime kilns, and smelt
dissolving tanks - are also regulated under this NSPS.
HAPs from pulping process equipment (predominantly methanol and smaller quantities of additional organic
compounds) are regulated by the US Environmental Protection Agency (USEPA) under the national emission
standards for hazardous air pollutants (NESHAPs) for pulp and paper production. This NESHAP also regulates
chlorinated compounds from bleaching processes at pulp and paper mills. A separate NESHAP regulates organic
HAPs (predominantly methanol, plus other organic compounds) and metallic HAPs (regulated through a PM
surrogate) from chemical recovery combustion sources at pulp mills. Federal NSPS regulate selected criteria
pollutants—nitrogen oxides (NOx), SO2, and PM—from industrial boilers, and the recently promulgated NESHAP
for industrial boilers and process heaters regulates HAPs from those sources.
Mill-specific criteria pollutant emission limits derived under USEPA's New Source Review and Prevention of
Significant Deterioration pre-construction permitting programs and emission limits from state regulations are
consolidated with these federal regulations in the title V operating permits of pulp and paper mills.
Over the past several decades, the pulp and paper industry has continually reduced its environmental impact by
increasing the use of recycled paper, improving energy efficiency, and making capital investments for effective
compliance with regulations. However, as noted in a 2009 document prepared by NCASI on the trade-offs and
benefits accompanying NOx and SO2 control (NCASI, 2009a), lingering environmental concerns associated with
emissions of NOx and SO2 have prompted continued pressure forfurther emissions reductions. These pollutants
originate as products of combustion that accompany power generation and the processing of pulping chemicals.
The NCASI report noted that measures have been taken in North America over the last 25 years to reduce
atmospheric emissions of NOx and SO2 where levels contributed to impaired environmental quality, as well as in
response to the aforementioned government-mandated performance standards. Nitrogen oxides and SO2
together have been implicated in adverse respiratory effects where certain thresholds are exceeded, as well as
acidic deposition thought to be of consequence to vegetation, soils and surface waters. Nitrogen oxide emissions
are also known to contribute to ozone formation and deposition-related eutrophication of surface waters. Most
recently, NOx and SO2 emissions are being scrutinized because of their role in the formation of fine PM (PM2.5),
which is an emerging health concern and a contributor to visibility impairment in certain geographic settings
(NCASI, 2009a).
GHGs are another source of concern for a number of industries, including the pulp and paper sector. Greenhouse
gas emissions from the pulp and paper sector are predominantly carbon dioxide (CO2), with smaller amounts of
methane (CH4) and nitrous oxide (N2O). The majority of the CO2 emissions from the pulp and paper industry are
biogenic CO2 emissions derived from the combustion of biomass fuels (e.g., bark and other wood residuals, black
liquor) that are generated onsite as a byproduct of the pulping process. Many pulp and paperfacilities generate
2-1
-------
over half of their energy needs from biomass fuels recovered from the pulp and paper production process
(USEPA, 2009).
A complex tool is needed to understand the technical and economic implications of applying process and
emission control technologies to reduce the emissions of NOx, S02, PM, and GHGs (particularly C02). To facilitate
the comprehensive analysis required to understand the complex interactions between economy and
environment, the USEPA has developed the Universal ISIS model. The Universal ISIS model has been populated
with data specific to the pulp and paper sector (Universal ISIS-PNP) to analyze the potential process and control
technologies for reducing these emissions from the pulp and paper industry. This chapter discusses the major
sources of NOx, SO2, PM, and CO2 in the pulp and paper industry and potential technologies for reducing the
emissions of these pollutants.
This chapter identifies emission reduction technologies, and, to the extent information is available in the
literature reviewed, an approximate percent reduction in emissions expected to be achieved with each
technology. When employing Universal ISIS-PNP for regulatory applications, users will be able to customize it
with updated control efficiencies developed through a more rigorous analysis of actual emissions test data. The
actual percent reduction that can be achieved with each technology depends on many factors, including process-
specific characteristics and baseline control strategies already in use.
2.2. Air Emissions Sources
Paper production is an energy intensive process. Power boilers at pulp and paper mills generate electricity and
process steam by combustion of fossil fuels and biomass. Some boilers fire so-called opportunity fuels such as
process gases, wastewater treatment sludges, etc. Recovery furnaces (sometimes referred to as recovery
boilers) at kraft pulp mills burn concentrated black liquor to recover cooking chemicals (specifically, Na2S) for
reuse in subsequent pulping cycles. While the primary purpose of the recovery furnace is to recover cooking
chemicals, the recovery furnace also produces heat used to generate steam and electricity for the mill. Kraft
pulp mills use lime kilns to convert lime mud from the white liquor clarifier to lime, which is used in the
causticizing process to recover additional pulping chemicals (specifically NaOH). Thermal oxidizers are used
mostly for the destruction of malodorous organic compounds and other non-condensable gases from the
pulping process. All of the above sources use fuel combustion for their operation and thus produce NOx, SO2
(depending on fuel used), and PM.
A recent survey study by NCASI, which estimated emissions from US pulp and paper mills (NCASI, 2012),
demonstrated that boilers are the dominant emission source of the NOx, S02, and PM emissions in the sector,
accounting for over 85 percent of the SO2, almost 65 percent of the NOx, and over 40 percent of the PM
emissions, as shown in Table 2-1, below.
2-2
-------
Table 2-1. 2010 Emissions, 103 tons (NCASI, 2012)
Process Unit
NOx
S02
PMa
Boilers
124
205
17
Gas Turbines
2
-
-
Kraft Recovery Furnaces
55
29
12
Kraft Smelt Dissolving Tanks
-
1
6
Kraft Lime Kilns
8
2
4
Kraft Thermal Oxidizers
1
1
<1
Sulfite Pulp Mills
3
1
<1
Semi-Chemical Pulp Mills
1
<1
<1
TOTAL
194
239
39
a Filterable PM only.
Recovery furnaces and lime kilns are also major emission sources of these pollutants, together accounting for
over 10 percent of the SO2, over 30 percent of the NOx, and over 40 percent of the PM emissions in the sector.
Compared to emissions from boilers, kraft recovery furnaces, and kraft lime kilns, emissions from sulfite and
semi-chemical mills (notably the chemical recovery combustion sources at these mills) are minimal, due to the
small numbers of these mills.
Greenhouse gas emissions from the pulp and paper source category are predominantly CO2 with smaller
amounts of CH4 and N2O. Fuel combustion is by far the largest source of GHG emissions emitted directly from
pulp and paper mill operations. Other non-energy-related sources of GHG emissions from pulp and paper mills
include use of carbonate-containing chemicals and CH4 releases from industrial wastewater treatment and
landfills. Table 2-2 summarizes the relative magnitude of nationwide GHG emissions (in million metric tons of
CO2 equivalents per year) reported to be emitted directly from stationary sources in the pulp and paper
manufacturing sector in 2004 (USEPA, 2010a).
Table 2-2. Nationwide GHG Emissions from the Pulp and Paper Manufacturing Industry
Emission Source
Million metric tons of COje per year
Direct emissions associated with fuel combustion (excluding biomass C02)
57.7
Wastewater treatment plant CH4 releases
0.4
Forest products industry landfills
2.2
Use of carbonate make-up chemicals and flue gas desulfurization chemicals
0.39
Direct emissions of C02 from biomass fuel combustion (biogenic)
113
Note: In addition to GHG emissions directly from each pulp and paper plant site, there are indirect GHG emissions associated with off-
site generation of steam and electricity that are purchased by or transferred to the mill. Indirect emissions have not been incorporated
into the current version of the Universal ISIS-PN P and are not discussed further in this document.
Biogenic CO2 emissions are of unique importance for the pulp and paper industry considering that the industry
satisfies much of its energy requirements by burning large quantities of biomass fuels. Biogenic CO2 emissions
2-3
-------
result primarily from combustion of spent pulping liquor in chemical recovery furnaces and from combustion of
woody biomass and other biogenic fuels in boilers and other combustion units. Biomass fuels have typically been
considered to be carbon neutral (i.e., treated as zero emissions) due to their short-term renewable life cycle for
purposes of emissions inventories; however, accounting methodsfor biogenic CO2 emissions are currently under
review by the USEPA. Once developed, USEPA's accounting methodology for biogenic CO2 emissions could be
applied in different policy contexts that are yet to be determined. Given that it is unknown at this time how the
biogenic CO2 accounting methodology will affect future emission reduction policies, Universal ISIS-PNP considers
the two extremes: (1) biogenic CO2 emissions could be considered as zero under policies analyzed in Universal
ISIS, or (2) biogenic CO2 emissions could be treated the same as any other CO2 emissions (e.g., derived from
fossil fuel combustion). A third and in-between scenario is that biogenic CO2 emissions could be discounted
based on regional or biomass feedstock-specific biogenic accounting factors that might place biogenic emissions
somewhere between zero and theirfull value.
Recent estimates of pulp and paper sector GHG emissions (excluding biogenic CO2 emissions) from USEPA's GHG
Reporting Program are presented in Table 2-3 below (USEPA, 2013).
Table 2-3. Pulp and Paper Sector — GHG Emissions Reported to the GHG Reporting Program for 2012
Emissions by GHG
Reporting year 2012
million metric tons of CC>2e per year*
Carbon dioxide (C02)
39
Methane (CH4)
0.9
Nitrous oxide (N20)
2
Total emissions (C02e)
42
* Biogenic C02 emissions are not included in emission totals provided above. Emissions from the industrial wastewater treatment and
landfills are not included in Table 2-3. Biogenic C02from the pulp and paper sector emissions were reported to be 121 million
metric tons in 2012. The global warming potential factors used to arrive at the totals in Table 2-3 were 1 for C02, 21 for CH4, and 310
for N2O.
The emissions in Table 2-3 are presented in CO2 equivalents (CC^e), which are derived by multiplying each GHG
by its respective global warming potential factor to place emissions on a common CC^e basis. Table 2-3 shows
that CChe emissions from fossil fuel combustion represent the majority of GHG emissions for the pulp and paper
sector. Methane and N2O from fossil fuel combustion are usually very small compared to CO2 emissions, even
after conversion to CChe. Thus, CO2 emissions represent the largest potential for GHG emission reductions in
the pulp and paper industry and, therefore, are the focus of GHG included in the Universal ISIS-PNP.
Further analysis of the 2012 GHG Reporting Program non-biogenic CO2 emissions data reveals that emissions
from boilers and pulp production (e.g., chemical recovery furnaces and lime kilns) represent the majority (95 %)
of the non-biogenic CO2 emissions from the pulp and paper industry. As shown in Figure 2-1, combustion
turbines, process heaters, incinerator control devices (used to combust non-condensable gases [NCGs] for HAPs,
volatile organic compounds [VOCs], and for TRS emissions control), and reciprocating internal combustion
engines comprise less than 5 percent of the combustion-related CO2 emissions. Figure 2-2 shows that, if biogenic
CO2 emissions were to be considered, then pulp production and boilers would account for nearly 99 percent of
the CO2 emissions. The Universal ISIS-PNP focuses on CO2 emissions from boilers, chemical recovery furnaces,
and lime kilns because these are the predominant GHG emission sources in the pulp and paper industry.
2-4
-------
1%_ 0.2%
0.0003%
I Boilers - 81%
I Pulp production (chemical
recovery and lime kilns) - 14%
Combustion turbines - 3%
I Process heaters - 1%
Incinerator control devices - 0.2%
Reciprocating internal combustion
engines - 0.0003%
Figure 2-1. Stationary Combustion and Pulp Production Sources of Non-Biogenic CO2 Emissions
0.04%
0.0001%
r
1
r
43%
}
i
56% 1
I
A
I Pulp production (chemical
recovery and lime kilns) - 56%
I Boilers-43%
Combustion turbines -1%
I Process heaters - 0.3%
Incinerator control devices -0.04%
Reciprocating internal
combustion engines -0.0001%
Figure 2-2. Stationary Combustion and Pulp Production Sources of Total CO2 (biogenic and non-biogenic CO2)
2.3. Boilers
2.3.1. Boiler Design and Fuels
The pulp and paper sector uses power boilers (in addition to recovery furnaces) to produce the steam and
electricity needed for the pulp and paper manufacturing process. According to NCASI study (NCASI, 2009a), the
pulp and paper sector uses nearly 1,000 of these auxiliary power boilers, with the following attributes:
• Approximately 30 percent of these boilers are larger than 250 million Btu (MMBtu) per hour. Less than 20
of these boilers are larger than 1000 MMBtu/h. The largest boiler is 1400 MMBtu/h.
2-5
-------
• Approximately 50 percent of the sector's power boilers were installed prior to 1970. Nearly 30 percent were
installed between 1971 and 1990. Approximately 20 percent were installed in 1991 or later.
The pulp and paper sector uses boilers with a variety of designs. Considering the design of heat exchange
between combustion gases and water, boilers can be of watertube orfiretube design. In the watertube design,
the fuel is combusted in a central chamber and the combustion gases transfer heat to water circulating in metal
tubes through radiation and convection. In the firetube design, water is stored in the main chamber of a boiler
and combustion gases flow through metal tubes within the body of the boiler, allowing for heat to be transferred
by conduction from the metal tubes to the surrounding water.
Boilers such as those described above for the pulp and paper sector have also been designed to operate with a
variety of fuels. The fuel mix for boilers for 1990 and 2010 is shown in Figure 2-3 in terms of the percentage of
total heat input. Coal, natural gas, wood, and residual oil are the primary fuels burned. The use of residual oil
has decreased significantly from 1990 to 2010. The heat input from residual oil has been replaced by the heat
input from wood. Boilers are commonly configured to burn multiple fuels to ensure that steam demands can be
met at the most favorable fuel cost (NCASI, 2009a).
Figure 2-3. Fuels Used by Boilers in Pulp and Paper Sector by Fleat Input (NCASI, 2012)
Coal-fired boilers most often use pulverized fuel and thus are known as pulverized-coal (PC) boilers. PC boilers
are used in large industrial units. Smaller industrial units use stoker-fired boilers. In PC boilers, coal is pulverized
to very small particle size in pulverizers or mills. These small coal particles are then blown with air into the boiler
where they are burned in suspension. Heat is transferred from the combustion gases to watertubes on the walls
of the boiler. PC boilers may be characterized by the burner configuration (wall, tangential, cyclone) and whether
the bottom ash exits the boiler in solid or molten state (dry bottom vs. wet bottom). Another type of coal-fired
boiler is a stoker boiler (stoker). In a stoker, the fuel is combusted in thin layers on top of a grate. Heat is
transferred from the combustion gases to watertubes on the walls of the boiler. Depending on how coal is
delivered to the grate, the stoker may be a spreader stoker (coal spread above the grate) or an underfeed stoker
(coal pushed into the bottom of the fuel bed). Other less common stoker types include traveling-grate, chain-
grate, and vibrating-grate.
Residual Oil, 1%
Wood, 44%
Wood, 33%
Natural Gas, 27%
Natural Gas, 27%
1990
2010
2-6
-------
Natural gas-fired boilers are typically smaller than coal-fired boilers and most often are package boilers. Based
on information from USEPA's AP-42 section on natural gas combustion (USEPA, 1998a), package boilers are
constructed off-site and shipped to the location where they are needed. While the heat input levels of packaged
units may range up to 250 MMBtu/h, the physical size of these units is constrained by shipping considerations.
The units generally have heat input levels less than 100 MMBtu/h. Package units are always wall-fired units with
one or more individual burners. Given the size limitations imposed on package boilers, they have limited
operational flexibility and cannot feasibly incorporate some NOx control options. Another type of natural gas-
fired boiler is a field-erected boiler. Field erected boilers are constructed onsite and comprise the larger sized
watertube boilers. Generally, boilers with heat input levels greater than 100 MMBtu/h are field-erected. Field-
erected units usually have multiple burners and, given the customized nature of their construction, also have
greater operational flexibility and NOx control options. Field-erected units can also be further categorized as
wall-fired ortangential-fired. Wall-fired units are characterized by multiple individual burners located on a single
wall or on opposing walls of the furnace, while tangential units have several rows of air and fuel nozzles located
in each of the four corners of the boiler (USEPA, 1998a).
Residual oil-fired boilers typically use Number 6 fuel oil or other heavy fuel oil. These oil-fired boilers are available
as package or field-erected units (USEPA, 2010b). In general, field-erected boilers are much more common than
package units in the boiler size category above 100 MMBtu/h input capacity, whereas below this capacity, the
boilers are usually package units. Field-erected boilers may be normal-fired ortangential-fired (NCASI, 2004).
Based on information from USEPA's AP-42 section on wood combustion (USEPA, 2003), wood or wood waste
(hog fuel) boilers are typically grate fired, with a spreader stoker employed for wood-fired boilers with a steam
generation rate larger than 100,000 Ib/h. In this boiler type, wood enters the furnace through a fuel chute and
is spread either pneumatically or mechanically across the furnace, where small pieces of the fuel burn while in
suspension. Simultaneously, larger pieces of fuel are spread in a thin even bed on a stationary or moving grate.
This type of boiler has a fast response to load changes, has improved combustion control, and can be operated
with multiple fuels. Natural gas, oil, and/or coal are often fired in spreader stoker boilers as auxiliary fuels. The
fossil fuels are fired to maintain constant steam production when the wood residue moisture content or mass
rate fluctuates and/or to provide more steam than can be generated from the residue supply alone. Although
spreader stokers are the most common stokers among larger wood-fired boilers, overfeed and underfeed
stokers are also utilized for smaller units. Dutch ovens and fuel cell ovens are two other grate-fired units used in
smaller operations. A later innovation in wood firing is the fluidized bed combustion boiler. A fluidized bed
consists of inert particles through which air is blown so that the bed behaves as a fluid. Wood residue enters in
the space above the bed and burns both in suspension and in the bed. Because of the large thermal mass
represented by the hot inert bed particles, fluidized beds can handle fuels with moisture content up to near 70
percent (total basis). Fluidized beds can also handle dirty fuels (up to 30 % inert material) (USEPA, 2003). Despite
their advantages, fluidized bed boilers represent only a small fraction of the population of boilers used in the
pulp and paper industry.
To address the complexity of the design-fuel matrix, boilers were grouped by the type of fuel used. This approach
will be used throughout the chapter to describe emissions from boilers and air pollution control technologies
applicable to the sector and in Universal ISIS modeling of technology application scenario.
2-7
-------
2,3.2, 5
The most important determinant of NOx, SO2, and PM emissions from boilers is the choice of fuel (NCASI, 2009a).
As noted in Figure 2-3, at pulp and paper mills in 2010, wood fuels accounted for 44 percent of the total fuel
heat input to boilers, followed by coal (28 %), natural gas (27 %) and fuel oil (1 %). Wood is most often burned
in combination with fossil fuels in these boilers (NCASI, 2009a). A comparison of the relative nitrogen and sulfur
content of various fuels is shown in Table 2-4.
Table 2-4. Relative Nitrogen and Sulfur Content of Fuels (NCASI, 2009a)
Fuel
Nitrogen, %
Sulfur, %
Natural Gas
Insignificant
Insignificant
Distillate Oil
0.05 or less
0.05 or less
Residual Oil
0.1 to 1.0
0.3 to 3.0
Coal
0.5 to 2.0
0.4 to 4.0
Bark and Wood Residue
0.1 to 0.4
0.2 or less
Also influential on NOx and SO2 emissions are features of the boiler's design (type, size) and the combustion
conditions under which it can be operated (boiler load, firing conditions) (NCASI, 2009a).
2.3.2.1.	Boiler NOx Emissions
The principal sources of NOx emissions from boilers are "thermal" NOx (formed from the thermal conversion of
nitrogen in the combustion air) and "fuel" NOx (formed from the nitrogen in the fuel) (NCASI, 2009a). Based on
information from an NCASI technical bulletin on criteria pollutant emissions from pulp and paper mills (NCASI
2004), the principal mechanism of NOx formation in natural gas combustion is the thermal NOx mechanism,
while NOx emissions from residual oil combustion arise from both fuel NOx and from thermal NOx. Fuel NOx can
account for 60 to 80 percent of the total NOx formation in residual oil combustion (NCASI, 2004).
NOx emissions from coal combustion (thermal and fuel NOx) are considerably higher than the NOx emissions
from gas or oil. Fuel NOx can account for up to 80 percent of the total NOx formed. Coal nitrogen content ranges
from 0.5 to 2 percent. Emissions of NOx are highest for cyclone boilers, followed by pulverized coal, stokers, and
mass feed units (NCASI, 2004).
Nitrogen oxide emissions from wood combustion are mainly the result of fuel NOx, with bark nitrogen contents
typically ranging from 0.1 to 0.2 percent. Average NOx emissions from wood combustion in typical pulp mill
boilers are lower than average NOx emissions from coal or residual oil combustion and slightly higher than
average NOx emissions from natural gas burning (NCASI, 2004).
2.3.2.2,	Boiler SO2 Emissions
Sulfur dioxide emissions are driven by fuel sulfur content, which is highest in coal and negligible in natural gas
(NCASI, 2009a). The average sulfur content of coal used in pulp and paper boilers was 1.27 percent for coals
used in 2010 (NCASI, 2012). Small amounts of other sulfur-containing fuels are burned in boilers, including tire-
derived fuel (TDF) and petroleum coke. TDF sulfur content is normally about 1.5 percent. Petroleum coke sulfur
content ranges from 4 to 6 percent.
2-8
-------
The incineration of NCG streams containing TRS in mill combustion devices creates the potential for emissions
of SO2 (NCASI, 2009a). The potential for SO2 emissions from this practice is relatively small compared with overall
mill emissions and varies with the combustion devices chosen (boiler, lime kiln, recovery furnace, or stand-alone
incineration device). Power boilers are the most versatile of the combustion devices used to incinerate NCG.
Approximately one-third of kraft mill power boilers are used to manage TRS gas streams because the relatively
large size of boilers accommodates high-volume, low-concentration and low-volume, high-concentration gas
streams. There is a potential increase of boiler SO2 emissions. However, SO2 can be absorbed by the alkaline
dust in wood and combination fuel boilers (NCASI, 2009a).
2.3.2.3.	Boiler PM Emissions
The determinants of PM emissions from pulp and paper boilers depend on a variety of factors, as outlined below
based on PM emissions information for each boiler type from USEPA's AP-42.
Because natural gas is a gaseous fuel, filterable PM emissions from natural gas boilers are typically low.
Particulate matter in natural gas combustion is usually higher molecular weight hydrocarbons that have not
been fully combusted. Increased PM emissions may result from poor air/fuel mixing or maintenance problems
(USEPA, 1998a).
PM emissions from residual oil burning are related to the oil sulfur content. Boiler load can also affect filterable
PM emissions in units firing residual oil, with low load conditions reducing emissions by 30 to 40 percent from
larger boilers and as much as 60 percent from smaller boilers. Under very low load conditions, proper
combustion conditions may be difficult to maintain, and PM emissions may increase significantly (USEPA, 2010b).
In coal-fired boilers, PM composition and emission levels are a complex function of boiler firing configuration,
boiler operation, pollution control equipment, and coal properties. Uncontrolled PM emissions from coal-fired
boilers include the ash from combustion of the fuel as well as unburned carbon resulting from incomplete
combustion. In pulverized coal systems, combustion is almost complete; thus, the PM emitted is composed
primarily of inorganic ash residues. Coal ash may either settle out in the boiler (bottom ash) or be entrained in
the flue gas (fly ash). The distribution of ash between the bottom ash and fly ash fractions directly affects the
PM emission rate and depends on the boiler firing method and furnace type (wet or dry bottom). Boiler load
also affects the PM emissions, as decreasing load tends to reduce PM emissions. However, the magnitude of the
reduction varies considerably depending on boiler type, fuel, and boiler operation (USEPA, 1998b).
In bark/wood combustion, PM emissions result from inorganic materials contained in the bark and wood itself
and from carbonaceous material resulting from incomplete combustion (NCASI, 2004).
2.3.2.4.	Boiler GHG Emissions
The paragraphs below summarize available information from USEPA's AP-42 on the sources of emissions of the
GHGs CO2, CH4, and N2O from boilers. These pollutants are all produced during combustion of natural gas,
residual oil, coal, and wood residues. Nearly all of the fuel carbon (99 % or more) is converted to CO2 during the
combustion process. This conversion is relatively independent of firing configuration. The majority of the fuel
carbon not converted to CO2 is due to incomplete combustion in the fuel stream. In natural gas and fuel oil
combustion, fuel carbon not converted to CO2 results in CH4, CO, and/or VOC emissions. In coal and wood
combustion, the majority of unconverted fuel carbon is entrained in bottom ash. Even in boilers operating with
poor combustion efficiency, the amount of CH4, CO, and VOC produced is insignificant compared to CO2 levels.
Carbon dioxide emissions from coal combustion vary with carbon content, and carbon content varies between
2-9
-------
the classes of bituminous and subbituminous coals. Further, carbon content also varies within each class of coal
based on the geographical location of the mine. Biogenic CO2 emitted from wood combustion has generally not
been counted as GHG emissions because of its role in the short-term CO2 cycle of the biosphere (USEPA, 1998a,
1998b, 2003, 2010b).
The formation of N2O during the combustion process is governed by a complex series of reactions, and its
formation is dependent upon many factors. Formation of N2O is minimized when combustion temperatures are
kept high (above 1475 °F), and excess air is kept to a minimum (less than 1 %). Nitrous oxide emissions for coal
combustion are not significant except for fluidized bed boilers, where the emissions are typically two orders of
magnitude higher than all other types of coal firing due to areas of low-temperature combustion in the fuel bed
(USEPA 1998a, 1998b, 2003, 2010b).
Methane emissions vary with the type of fuel and firing configuration, but are highest during periods of
incomplete combustion or low-temperature combustion such as the start-up or shut-down cycle for boilers.
Typically, conditions that favor format ion of N2O also favor emissions of CFU( USEPA, 1998a, 1998 b, 2003, 2010b).
2,3.3. Boiler Emission Reduction Strategies
2,3.3.1. Boiler NOx Reduction
As noted previously, NOx is formed in boilers mostly through the oxidation of nitrogen in the combustion air
(thermal NOx) and through oxidation of fuel nitrogen (fuel NOx). According to NCASI 2009a, the firing of natural
gas typifies the former, whereas the firing of coal and oil typifies the latter. Fuel NOx represents approximately
50 percent of the total uncontrolled emissions when firing residual oil and more than 80 percent when firing
coal (NCASI, 2009a).
NCASI 2009a indicated that fuel switching is an attractive option for reducing boiler NOx emissions, but
cautioned that its application cannot be considered in isolation from a host of site-specific factors of importance
to boiler performance, boiler integrity and overall emissions control capability. For example, biomass and wood
are favorable fuels from the standpoint of NOx emissions, but firing them has been observed to lead to
accelerated corrosion of boiler components. Fuel properties are best taken into account at the time of boiler
design (NCASI, 2009a).
Apart from choice of fuel, control technologies exist that can reduce boiler NOx emissions. These technologies
can be divided into primary and secondary control technologies. Primary control technologies seek to limit the
formation of thermal NOx by manipulation of combustion conditions, while secondary control technologies aim
to remove the NOx from the flue gas by treatment of flue gas in the post-combustion regions of the furnace
(NCASI, 2009a). Various approaches for NOx control are characterized in Table 2-5. According to NCASI 2009a,
the applicability of individual options and performance will depend upon boiler design and configuration, fuels
being burned, and the dynamic character of boiler loading. Greater opportunity for NOx reduction exists when
the capability is designed into newly constructed boilers as opposed to retrofitting existing boilers (NCASI,
2009a).
Table 2-5. Boiler NOx Control Technologies (NCASI, 2009a)
Control Option
Description
Performance
Application
Combustion Modifications
2-10
-------
Control Option
Description
Performance
Application
Low-NOx Burner
(LNB)
Burners designed to mix fuel and air
in a controlled pattern that sustains
local fuel-rich regions, keeps the
temperatures down and dissipates
heat quickly
Approximately 50 % NOx
reduction
Used in both gas/oil-fired and coal-fired
units. Elongated flame configuration
limits application in smaller boilers
Ultra Low-NOx
Burner (ULNB)
For gaseous fuel burners, ULNBs
often use air staging and internal flue
gas recirculation (FGR) (no external
ductwork needed), or they may
alternatively use lean-premixed
combustion with FGR for lower
emissions than possible with LNB
alone. ULNB is also a term used for
some second-generation coal-fired
LNBs that are installed in
combination with overfire air (OFA)
(Andover 2010).
In the range of 75 % NOx
reduction (Andover, 2010)
Used in gas-fired and coal-fired units
(Andover, 2010)
Flue Gas
Recirculation
(FGR)
Up to 20 % of the combustion flue
gas is brought into the combustion
zone, acting as a heat sink, lowering
combustion zone temperature
20 -30 % NOx reduction
Because only thermal NOx formation
can be controlled by this technique, it is
especially effective only in oil- and gas-
fired boilers. Most effective when used
in conjunction with air and/or fuel
staging. More adaptable to new designs
than as a retrofit application. Capital
intensity and high operating and
maintenance (O&M) costs are
prejudicial for use on industrial-scale
boilers
Overfire Air
(OFA)
Diversion of 10-20 % of combustion
air downstream of burners
15-30 % NOx reduction
More attractive for new units than
retrofit applications. May be used with
all fuels and most combustion systems;
Can decrease energy efficiency
Biased Burner
Firing
The furnace is divided into a lower,
fuel-rich zone and an upper fuel-lean
zone to complete the burnout
20 % NOx reduction
Proven only for oil/gas-fired utility
boilers
2-11
-------
Control Option
Description
Performance
Application
Low Excess Air
Reducing excess air in the
combustion flame zone reduces fuel
and thermal NOx formation
-	Dry bottom: 50-7 0% NOx
reduction
-	Wet bottom: 30-70 %
NOx reduction
-	Fluidized bed: no data
-	Traveling grate stoker:
35-50 % NOx reduction
-	Spreader grate stoker:
50-65 % NOx reduction
(USEPA 1999)
Limited by production of smoke, high
CO emissions, and increased fouling
and corrosion in boiler. Applied for
energy efficiency.
Fuel Staging
10-20 % of the total fuel input is
diverted to a second combustion
zone downstream of the primary
zone. Combustion of fuel in the fuel-
rich secondary zone reduces nitric
oxide (NO) formed in the primary
zone to nitrogen (N2). Low nitrogen-
containing fuels such as natural gas
and distillate oil are typically used for
reburning to minimize further NOx
formation.
Claims of NOx reductions
from 50-70 % when
combining this approach
with overfire air and flue
gas recirculation
Limited application in the US
Burners Out of
Service
In multiple burner systems, fuel flow
is blocked to upper burners allowing
only air to pass
50-70 % NOx reduction for
dry bottom boilers (USEPA
1999)
Useful in retrofit situations involving
suspension-fired coal and oil/gas-fired
boilers. Operational problems can
include soot/slag formation
Reduced Air
Preheat
Lowers the primary combustion zone
peak temperature through reduced
preheating of the combustion air
-	Dry bottom: 50-70 % NOx
reduction
-	Wet bottom: 30-70 %
NOx reduction
-	Fluidized bed: no data
-	Traveling grate stoker:
35-50 % NOx reduction
-	Spreader grate stoker:
50-65 % NOx reduction
(USEPA 1999)
Reduced air preheat lowers only
thermal NOx, and thus is economically
attractive only for natural gas and
distillate fuel oil combustion. The
energy penalty usually makes this
option unfavorable.
Steam & Water
Injection
Flame quenching by the addition of
steam or water in the combustion
zone
-	Drv bottom: 50-70 % NOx
reduction
-	Wet bottom: 30-70 %
NOx reduction
-	Traveling grate stoker:
35-50 % NOx reduction
-	Spreader grate stoker:
50-65 % NOx reduction
(USEPA 1999)
An effective control technology for
oil/gas-fired burners, but one with a
potentially significant energy penalty
Load Reduction
Reducing boiler capacity lowers
flame temperatures and reduces
thermal NOx formation
NOx reduction specific to
boiler capacity
Can cause improper fuel-air mixing
during combustion, creating carbon
monoxide (CO) and soot emissions
2-12
-------
Control Option
Description
Performance
Application
Post-Combustion/Flue Gas Treatments
Selective Non-
Catalytic
Reduction (SNCR)
Involves the injection of urea,
ammonium hydroxide, anhydrous
ammonia, or aqueous ammonia into
the furnace exit region where the
flue gas is in the range of 900 to
1,150 °C. Nitrogen oxide is reduced
to N2and water. Performance
affected by inlet NOx level,
temperature, mixing, residence time,
reagent-to-NOx ratio, and fuel sulfur
content.
NOx reduction as high as
60-70 %
A portion of the NO reduction (about 5
%) is due to formation of N20, a potent
GHG. Process complexity prompts
concern about ability to perform
adequately under changing load and
fuel conditions. Operating problems
include optimizing chemical addition to
prevent ammonia emissions in the flue
gas and, with higher sulfur fuels, salt
deposits on downstream components
that contribute to plugging and reduced
heat transfer.
Selective
Catalytic
Reduction (SCR)
NOx is reduced to N2 and water by
the injection of ammonia into the
flue gas at temperatures between
350 and 400 °C in the presence of a
catalyst. Performance is affected by
NOx level at SCR inlet, flue gas
temperature, ammonia-to-NOx ratio,
fuel sulfur content, gas flow rate, and
catalyst condition.
70-90% NOx reduction
A proven technology, but not often
applied to smaller industrial-scale
boilers. Major problems with SCR
processes include corrosion, formation
of solid ammonium sulfate, and
formation of salt deposits in high sulfur
oil-fired or coal-fired boilers that reduce
heat transfer efficiencies. Ammonia
slippage* is also a potential problem.
Catalysts lose activity over time due to
poisoning by trace metals or erosion by
fly ash.
Regenerative SCR
(RSCR)
RSCR applies SCR in combination
with regenerative thermal oxidizer
technology to more efficiently reheat
cleaned gas to SCR operating
temperatures than possible in
previous "tail end" SCR designs. RSCR
is a recently-developed technology
that has been used on biomass-fired
boilers (Andover, 2010).
>75 % NOx reduction
(Andover 2010)
Used on biomass boilers downstream of
PM removal devices to reduce NOx.
Therefore, RSCR is well suited for many
pulp and paper mills. RSCR has the
advantage of being installed near the
end of the process and requires less
fuel to reheat the gas than traditional
"tail end" SCR systems. As a result, RSCR
may be an attractive retrofit option.
RSCR may be limited by the available
space near the existing chimney
(Andover, 2010).
Low-
Temperature
Oxidation
(LoTOx)
LoTOx is a process whereby the NOx
compounds are oxidized to water-
soluble forms, which are
subsequently captured in a
downstream wet scrubber (Andover,
2010).
>90 % NOx reduction
(Andover, 2010)
This process can perform only in
combination with a downstream wet
scrubber (Andover, 2010).
* When ammonia passes through the SCR unreacted, it is known as "slippage." Slippage can result from over-injection into the gas
stream, catalyst degradation, or if the temperature is not high enough for the ammonia to react.
Note: all data from (NCASI, 2009a), except where otherwise indicated.
Thermal NOxformation is commonly controlled by reducing peak and average flame temperatures, an approach
contrary to measures typically employed to ensure complete fuel combustion. A compromise is therefore
exacted between effective combustion and NOx formation. Conversion of fuel-bound nitrogen is more
dependent upon fuel-air proportions than variations in combustion zone temperatures. NOx control involves a
2-13
-------
delicate balance of air distribution and combustion temperature control that invites a risk of combustion
inefficiency and potential release of pollutants associated with incomplete combustion (NCASI, 2009a).
Post-combustion flue gas controls involve chemical reduction of NOx to nitrogen (N2), entailing the injection of
ammonia-based compounds under suitable temperature conditions where flue gas exits the furnace (NCASI
2009a). Because of the relatively narrow temperature windows required and reaction chemistry sensitivity to
flue gas flow rates, NCASI concluded that these control options are ill-suited for application to industrial scale
boilers that are subject to highly variable loads and fuel combinations (NCASI, 2009a).
LNB/FGR. Low-NOx burners (LNBs), as well as FGR, are the most widely applied primary technologies for boiler
NOx reduction (NCASI, 2009a). LNB limits NOx formation by controlling the stoichiometry and temperature of
combustion. LNBs may use staged combustion to slow complete fuel-air mixing or lean-premixed combustion
(mostly for gas fuel). This staged combustion reduces both flame temperature and oxygen concentration during
some phases of combustion, lowering both thermal NOx and fuel NOx formation (NCASI, 2009a). NOx reductions
up to approximately 50 percent may be achieved by LNBs. The extent of reduction depends on fuel preparation
and local conditions in the furnace. Flue gas recirculation reduces thermal NOx formation by reducing peak
temperatures and limiting oxygen availability. Taken together, NCASI indicated that NOx reductions of 60 to 90
percent are achievable. Flowever, the report concluded that flue gas recirculation is better suited to new boilers
rather than retrofits, can reduce boiler heating capacity, and is difficult to justify economically for industrial-
scale boilers (NCASI, 2009a).
ULNB. Ultra low-NOx burners (ULNB) often use air staging and internal FGR (no external ductwork needed), or
they may alternatively use lean-premixed combustion with FGR for lower emissions than possible with LNB alone.
ULNB is also a term used for some second-generation coal-fired LNBs that are installed in combination with OFA.
ULNBs offer in the range of 75-percent reduction from uncontrolled levels (Andover, 2010).
OFA. Another commonly used primary technology is OFA, which is a form of air staging in which a fraction
(typically 10-20 %) of combustion air is injected downstream of the burner. OFA is often used in conjunction
with LNBs to increase NOx reduction by an additional 15 to 30 percent. Use of OFA can result in reduced boiler
efficiency manifested by increased CO concentration and loss on ignition in the flue gas.
SCR/SNCR. Two of the secondary technologies that could be used to reduce NOx emissions include selective non-
catalytic reduction (SNCR) and selective catalytic reduction (SCR). Both SNCR and SCR involve injection of a
reducing agent such as ammonia or urea into the flue gas under conditions where the reagent can react with
NOx to form N2 and water (NCASI 2009a). In SNCR, the reducing agent reacts with NOx at about 900 to 1150 °C,
while in SCR, the reduction reaction occurs at around 350 to 400 °C and for this reason requires a catalyst. The
catalyst is typically installed between the boiler economizer and air preheater (known as a hot side or high-dust
installation). Sometimes the SCR reactor can be placed after the air preheater (known as the low-dust SCR). The
catalyst needs to be replaced periodically because of its sensitivity to impurities in the flue gas, resulting in
catalyst poisoning and/or blinding. Catalysts have been found to remain active much longer in flue gas from the
combustion of natural gas than from the combustion of coal.
In SNCR, a stoichiometric excess of reducing agent is needed forthe reaction to proceed effectively, creating the
operational requirement to limit the so-called ammonia slip (unreacted agent that exits with the flue gas) that
can impact plume visibility and make fly ash difficult to dispose of. SNCR has been widely used on boilers and
2-14
-------
has a proven NOx reduction of over 60 percent. SCR has also been widely used and has achieved NOx reductions
in excess of 90 percent.
According to NCASI 2009a, the retrofitting of SCR on industrial boilers has been found to be difficult and costly,
indicating that, in retrofit application, capital costs are estimated to be 30 to 50 percent higher. Moreover, SCR
systems are not very tolerant of constantly changing conditions, as a stable window of operation is required for
optimum efficiency. Load swings make it particularly difficult to retrofit boilers with SCR or SCNR, as appropriate
temperature windows are hard to maintain. Urea- or ammonia-handling systems are an added complication for
boiler operations. In addition, associated salt deposition on downstream boiler components contributes to
plugging and reduced heat transfer efficiencies (NCASI, 2009a).
NCASI 2009a further noted that secondary emissions that can result with SNCR include such intermediate
reaction products as N2O, a potent GHG. Nitrous oxide levels have been observed to equal up to 4 percent of
the NOx reduction with ammonia injection, while urea injection yielded N2O levels up to 25 percent of the NOx
reduced. However, the report noted that SCR has been found to enhance mercury removal (NCASI, 2009a).
The use of SCRs is often limited by the available space to install the catalyst reactor at the correct temperature
that exists in the process, which may require significant changes to the existing equipment unless a "tail end"
SCR is installed, where the gas is reheated to the correct temperature. However, "tail end" SCR units are
unattractive due to the additional fuel necessary for reheating the gas. An alternative to a traditional "tail end"
SCR is a regenerative SCR (RSCR). RSCR applies SCR in combination with regenerative thermal oxidizertechnology
to more efficiently reheat cleaned gas to SCR operating temperatures than possible in a previous "tail end" SCR
designs. RSCR is a recently-developed technology that has been used on biomass-fired boilers downstream of
PM removal devices to reduce NOx by over 75 percent. Therefore, RSCR is well suited for many pulp and paper
mills. RSCR has the advantage of being installed near the end of the process and requires less fuel to reheat the
gas than traditional "tail end" SCR systems. As a result, it may be an attractive retrofit option. RSCR may be
limited by the available space near the existing chimney (Andover, 2010).
Multi-pollutant reduction. One example of this type of emission control involves the use of SCR followed by wet
flue gas desulfurization (FGD), which has gained credence as a potential means of reducing not only NOx and
SO2, but also mercury emissions (NCASI, 2009a). The contribution of SCR technology to mercury reduction comes
from the fact that SCRs have been shown to oxidize elemental mercury. Wet scrubbers, in turn, have been shown
to be effective in removing oxidized mercury (NCASI, 2009a). Another example involves the use of low
temperature oxidation (LoTOx), whereby the NOx compounds are oxidized to water-soluble forms that are
subsequently captured in a downstream wet scrubber. The oxidizer used to convert NOx in the LoTOx process
may also help remove SO2 by oxidizing it to sulfur trioxide (SO3). Nitrogen oxide emission reductions higher than
90 percent may be achievable using the LoTOx process. However, this process can perform only in combination
with a downstream wet scrubber (Andover, 2010).
Applicability of NOx control technologies. As noted above, the applicability of individual NOx control options and
performance will depend upon boiler design and configuration, fuels being burned, and the dynamic character
of boiler loading. The following paragraphs review the applicability of the primary and secondary NOx control
technologies for the different boiler types at pulp and paper mills, summarized in Table 2-6, below.
2-15
-------
Table 2-6. Applicability of NOx Control Technologies (Andover, 2010)
Boiler Type
Wood/Hog Fuel
Coal
Natural Gas
Residual Oil
LNB
No
Yes(50)b
Yes(50)
Yes (50)
ULNB
No
No
Yes (75)
No
OFA
Yes (25)
Yes(25)
No
Yes (25)
SNCR
Yes (50)
Yes(50)
Noc
Yes(25)
SCR
Yes (75)a
Yes(90)
Yes (90)d
Yes(90)
RSCR
Yes(75)
Yes(75)
Yes (75)d
Yes(75)
LoTOxe
Yes (90)
Yes(90)
Yes(90)
Yes (90)
Note: percent NOx reduction in parentheses.
aTail-end configuration.
b Pulverized coal only.
c Retrofit possible; not on new units.
d New units possible; not on retrofits.
e Requires downstream wet scrubber.
Wood and hog fuel boilers are typically grate-fired, possibly fluid- or bubbling bed-fired, and are not amenable
to traditional low-NOx burners. Wood and hog fuel boilers may use any of the post-combustion NOx control
methods described above. However, for SCR application, only a tail-end SCR configuration would be applicable
because of the need to avoid catalyst poisons. SNCR could be used and would be expected to provide
approximately 50-percent reduction of NOx-
RSCR has been used on biomass boilers downstream of PM removal devices to reduce NOx, so RSCR is well suited
for many pulp and paper mills. As noted previously, RSCR has the advantage of being installed near the end of
the process and requires less fuel to reheat the gas than traditional "tail end" SCR systems. As a result, RSCR
may be an attractive retrofit option. RSCR may be limited by the available space near the existing chimney,
however. SCR and RSCR are not likely to be used for retrofit of gas-fired boilers due to low NOx levels that are
achievable with combustion controls. However, SCR and RSCR are an option for new gas boiler installations.
Application of LoTOx would likely be limited because this process would require a downstream wet scrubber
(Andover, 2010).
Pulverized coal-fired boilers can use combustion modifications such as low-NOx burners and OFA. Grate-fired
boilers would not use low NOx burners and would instead use air staging similar to OFA. Any post-combustion
NOx control technology could be used on coal-fired boilers (pulverized orgrate), and the selection would depend
on the desired level of NOx control and on the size of the facility. SNCR would typically provide up to
approximately 50-percent reduction. Because smaller boilers (<100 MW) would typically be expected at pulp
and paper facilities, there might be space limitation and thus tail-end SCR would be likely for smaller boilers. For
utility-size coal-fired boilers, conventional high-dust SCR would be expected.
Natural gas boilers typically have their NOx emissions most effectively controlled with combustion modification
controls such as low-NOx burners or flue gas recirculation. As far as post-combustion controls, SNCR is not likely
to be used on natural gas boilers with low-NOx burners. SCR would likely be installed on most new facilities.
Retrofit of the SCR into the required temperature zone on natural gas boilers could be difficult and might require
2-16
-------
either reheat or the use of a low-temperature SCR. However, SCR would not likely be a good retrofit candidate
since combustion modifications would bring NOx emissions to low levels on their own.
Residual oil is burned in numerous boilers at pulp and paper mills. They typically can use LNB and OFA for
combustion modifications and may use any of the post-combustion control methods. If an SCR system is installed,
a tail-end SCR may be necessary for these boilers for the reasons explained earlier for coal-fired boilers.
2,3,3,2, Boiler SO2 Reduction
Based on information from NCASI, fuel switching is an attractive option for reducing boiler SO2 emissions, but
the report cautioned that its application cannot be considered in isolation from a host of site-specific factors of
importance to boiler performance, boiler integrity and overall emissions control capability. The report stated
that switching to lower sulfur fuels can be an effective way to reduce SO2 emissions but pointed out that lower
sulfur fuels are typically more expensive and indicated that there is a question of compatibility with the design
of the existing boiler system and related equipment. The report further cautioned that fuel changes may also
compromise boiler efficiency and emissions control capability. To illustrate, the report cited an example of a
boiler switching from (1) eastern bituminous coal, with a high heat value and low ash content, to (2) a low-sulfur
western sub-bituminous coal with a lower heating value and high ash content. Though such a change may be
beneficial for reducing SO2 emissions, the report noted that it comes with the following potentially adverse
effects (NCASI, 2009a):
•	Flame stability impacts consequential to boiler efficiency and pollutant emissions
•	Diminished energy efficiency due to deposition and slagging on heat transfer surfaces
•	Increased ash loading
•	Unsatisfactory performance of emissions control equipment.
According to the report, natural gas is recognized as a clean burning fuel, but its higher hydrogen content yields
water vapor during combustion that contributes to greater heat loss out the stack. The report noted that
biomass and wood are favorable fuels from the standpoint of SO2 emissions, but firing these fuels has been
observed to lead to accelerated corrosion of boiler components. The report suggested that fuel properties are
best taken into account at the time of boiler design (NCASI, 2009a).
NCASI 2009a noted that post-combustion FGD techniques can be used to remove SChformed during combustion
of sulfur-bearing fuels (e.g., coal). FGD involves injection of an alkaline sorbent into the flue gas stream that
reacts with SO2 to form liquid or solid sulfur-bearing compounds that are subsequently separated; SO2 FGD
scrubber systems are characterized as either wet, dry, or semi-dry, as well as non-regenerable or regenerable in
terms of whether the end products have viable commercial use (NCASI, 2009a). Attributes of various SO2 control
technologies are summarized in Table 2-7.
2-17
-------
Table 2-7. Boiler SO2 Control Technologies (NCASI, 2009a)
Control Option
Description
Performance
Application
Wet Systems
Lime/Limestone Wet
Scrubbing (LWS)
Aqueous slurry of the sorbent is injected
into the flue gas, saturating the gas
stream. Sulfur dioxide dissolves into
slurry droplets and reacts with alkaline
particles. The slurry falls to the bottom
of the reactor, is collected, and sent to a
reaction tank to complete conversion to
a neutral salt.
80 to 90 % S02
removal with
limestone; up to
95 % removal
with lime
Wet systems are applicable to high sulfur
fuels and produce a wet sludge
byproduct requiring management and
disposal. Though high in capital and
operating cost, wet limestone scrubbing
is the preferred process for coal-fired
electric utility plants.
Sodium Carbonate,
Hydroxide or
Bicarbonate Wet
Scrubbing
80% to 98%
reduction
High reagent cost a disadvantage
Magnesium
Oxide/Hydroxide Wet
Scrubbing
80% to 95+%
reduction
Sorbent can be regenerated
Dual Alkali Wet
Scrubbing
90% to 96%
reduction
Uses lime to regenerate sodium-based
scrubbing liquid
Semi-Dry Systems (Spray Dryer Absorber [SDA])
Calcium Hydroxide
Slurry Sorbent
As with wet systems, aqueous sorbent
slurry is injected into the flue gas stream.
The sorbent is more concentrated in
semi-dry system slurries, however. Hot
flue gas evaporates water in the slurry,
but sufficient water remains on the solid
sorbent to enhance S02 removal. The
resulting dried waste product is
subsequently captured with a standard
particulate collection device.
70 % to 90 % S02
reduction
Applicable to low- and medium-sulfur
fuels; produces a dry residual byproduct
that is less difficult to manage than wet
residuals. Performance is sensitive to
operating conditions due to potential for
wet solids to deposit on the absorber
and downstream equipment. High
temperatures and high SO2
concentrations degrade performance.
Typical applications are utility and
industrial boilers burning low to medium
sulfur coal and requiring 80 % S02
control.
Dry Systems
Dry Calcium
Carbonate/Hydrate
Injected in Upper
Furnace Cavity
Powdered sorbent is injected directly
into the furnace. The waste product is
removed with standard particulate
control equipment.
50-60 % S02
reduction
Even distribution of sorbent and
adequate residence time within narrow
temperature bands is critical for high SO2
removal. Dry systems are less costly than
wet systems, use less space, and are
thought more suitable for retrofit
applications. The technique is viewed as
an emerging technology for medium-to-
small industrial boiler applications.
Dry Sorbent Injection
into Ductwork
Powdered sorbent is injected directly
into downstream ductwork. Water can
be injected to enhance S02 removal. The
waste product is removed with standard
particulate control equipment.
50-80 % S02
reduction with
sodium-based
sorbent.
Wet FGD. Wet systems, the most commonly employed technique, achieve the greatest removals, with SO2
reductions of 95 percent and more (NCASI 2009a). The wet scrubbing process most commonly used to treat
boilerflue gas is limestone wet scrubbing (LWS). Occasionally, lime wet scrubbing or sodium wet scrubbing may
be used. LWS uses a low cost reagent and can capture up to 90 percent of SO2 (depending upon inlet SO2 levels).
In addition, LWS generates a byproduct that can be disposed or reused. LWS is most often used for high SO2
2-18
-------
concentration applications. Lime wet scrubbers are similar in operation to LWS. The slurry is more reactive than
limestone slurry, allowing for the same level of SO2 removal in a smaller scrubber (lower capital costs) as LWS,
but lime is more expensive than limestone. The use of dolomitic lime allows for further decrease in the size of
the wet scrubber compared to conventional lime reagent. When sodium is used as a reagent for wet scrubbing
of SO2, no solid waste is produced. The byproduct from the sodium wet scrubber can be converted to Na2S in
the recovery furnace, and this conversion may create the potential to mitigate the cost of chemicals used in the
pulping process.
NCASI 2009a noted that wet FGD using lime/limestone is used primarily for reducing SO2 emissions for large
electric utility boilers and concluded that the technology cannot be cost-justified for industrial-scale boilers. The
report cited a cost survey carried out by the EUCG (formerly known as the Electric Utility Cost Group) that
documented the sensitivity of cost to boiler size. As shown in Figure 2-4, costs for FGD systems for boilers smaller
than 300 MW are nearly double the costs for boilers greater than 300 MW. Most boilers in the pulp and paper
sector are significantly smaller than 300 MW, with the average size being equivalent to roughly 25 MW. These
small boilers would be subject to disproportionate costs were they to adopt this control technique (NCASI,
2009a).
<300	300-600 600-900	>900	ALL
PLANT SIZE RANGE (MW)
Figure 2-4. FGD-Only Costs among 49 FGD Systems (NCASI 2009a)
Space availability is another aspect that can skew the costs of FGD system installation (NCASI, 2009a). Pulp and
paper mills house a vast array of large-scale process equipment concentrated in a relatively small footprint.
Accommodating an FGD system would incur disproportionate construction costs. Such space constraints might
favor a dry FGD system. Dry systems have been characterized as an emerging technology for industrial-scale
boilers. Flowever, the dynamic nature of mill boiler loadings would jeopardize performance, given the sensitivity
of dry systems to operating conditions (NCASI, 2009a).
The report noted that, within the pulp and paper industry there are numerous fluidized-bed boilers with lime
injection for SO2 removal, plus many more boilers with wet control devices (venturi scrubbers, wet electrostatic
2-19
-------
precipitators [ESPs], spray towers) and alkali addition for SO2 removal. There are no lime/limestone wet FGD
systems of the type that dominate coal-fired electric utility boiler systems (NCASI, 2009a).
Semi-dry FGD. Both wet and semi-dry FGD approaches impose water demand ranging from 0.5 to 1.5 tons of
water per ton of coal burned (NCASI, 2009a). Fleating and evaporation of that water also impose a significant
energy demand. The need to reheat flue gas to preserve plume buoyancy poses an additional drain. The
electrical energy required to drive process equipment has been estimated to range from 1 to 2.5 percent of
boiler capacity. The report noted that schemes exist to regenerate the chemical absorbent, but stated that the
regeneration schemes are very energy-intensive. The report also noted that once-through systems are most
common, but stated that the once-through systems generate a large quantity of solid wastes. The report further
noted that, while the accumulation of metals, including mercury, in wastewaters and sludges of FGD systems is
of benefit to air emissions, it is problematic with regard to the management of those waste streams. Removal
of mercury from flue gas, however, is a co-benefit (NCASI, 2009a).
The report pointed out the following from the comparisons of wet and semi-dry approaches (NCASI, 2009a):
•	The non-air quality environmental impacts and negative energy impacts are significantly greater for the wet
FGD control technology, since it generates a visible plume, consumes more water, generates a wastewater
stream requiring treatment and disposal, generates slightly more solid byproducts for landfill, and because
the wet FGD requires significantly more auxiliary power consumption during operation.
•	Compared to wet lime/limestone scrubbing technology, the spray dryer absorber (SDA) has the reported
advantages of fewer major equipment items and thus lower capital cost, high reliability, lower space
requirements, lower potential for corrosion, potential for lower energy consumption, absence of a
wastewater stream, lower water consumption, and less sensitive and simpler process chemistry.
Dry FGD. Dry scrubbers typically do not achieve the SO2 reduction levels associated with their wetter
counterparts but stated that the technology does offer other relative advantages. Specifically, the dry scrubbers
have significantly lower capital and operating costs because they are simpler, demand less water, and involve
less complex waste disposal (NCASI, 2009a).
Multi-pollutant reduction. One example of this type of emission control involves the use of SCR followed by wet
FGD, which has gained credence as a potential means of reducing not only NOx and SO2 but also mercury
emissions (NCASI, 2009a). The contribution of SCR technology to mercury reduction is that SCRs have been
shown to oxidize elemental mercury; wet scrubbers, in turn, have been shown to be effective in removing
oxidized mercury (NCASI, 2009a). Another example involves the use of low temperature oxidation (LoTOx),
where the NOx compounds are oxidized to water-soluble forms that are subsequently captured in a downstream
wet scrubber. The oxidizer used to convert NOx in the LoTOx process may also help remove SO2 by oxidizing it
to SO3. Nitrogen oxide emission reductions higher than 90 percent may be achievable using the LoTOx process.
Flowever, this process can perform only in combination with a downstream wet scrubber (Andover, 2010).
2,3,3,3, Boiler PM Reduction
The following paragraphs summarize available information on various control technologies for reducing PM
emissions, based on information provided in USEPA's AP-42.
2-20
-------
As with NOx and SO2, fuel switching is an attractive option for reducing boiler PM emissions. For example, PM
will generally be reduced when a lighter grade of fuel oil is burned. Fuel alteration of heavy oil by mixing with
water and an emulsifying agent has also reduced PM emissions significantly in controlled tests (USEPA, 2010b).
Apart from choice of fuel, the principal PM control techniques for industrial size boilers are post-combustion
methods, including ESP, fabric filter (or baghouse), wet scrubber, or mechanical collector (USEPA, 1998b, 2003,
2010b). Attributes of these PM control technologies are summarized in Table 2-8 and discussed in the
paragraphs below.
ESP. Electrostatic precipitation technology is applicable to a variety of coal combustion sources. Because of their
modular design, ESPs can be applied to a wide range of system sizes and should have no adverse effect on
combustion system performance. The operating parameters that influence ESP performance include fly ash
mass loading, particle size distribution, fly ash electrical resistivity, and precipitator voltage and current. Other
factors that determine ESP collection efficiency are collection plate area, gas flow velocity, and cleaning cycle.
Data for ESPs applied to coal-fired sources showfractional collection efficiencies greaterthan 99 percent forfine
(less than 0.1 micrometer) and coarse (greater than 10 micrometers) particles. These data show a reduction in
collection efficiency for particle diameters between 0.1 and 10 micrometers (USEPA, 1998b).
Atmospheric fluidized bed combustion (AFBC) boilers may tax conventional particulate control systems. The
particulate mass concentration exiting AFBC boilers is typically 2 to 4 times higher than the particulate mass
concentration exiting pulverized coal boilers. AFBC particles are also, on average, smaller in size and irregularly
shaped with higher surface area and porosity relative to pulverized coal ash. The effect is a higher pressure drop.
The AFBC ash is more difficult to collect in ESPs than pulverized coal ash because AFBC ash has a higher electrical
resistivity, and the use of multiclones for recycling, inherent with the AFBC process, tends to reduce the exit gas
stream particulate size (USEPA, 1998b).
Electrostatic precipitators are commonly used in oil-fired power plants. Older precipitators, usually small,
typically remove 40 to 60 percent of the emitted PM. Because of the low ash content of the oil, greater collection
efficiency may not be required. Currently, new or rebuilt ESPs can achieve collection efficiencies of up to 90
percent (USEPA, 2010b).
Electrostatic precipitators are employed with wood-fired boilers when collection efficiencies above 90 percent
are required. When applied to wood-fired boilers, ESPs are often used downstream of mechanical collector pre-
cleaners that remove larger-sized particles. Collection efficiencies of 90 to 99 percent for PM have been
observed for ESPs operating on wood-fired boilers (USEPA, 2003).
A variation of the ESP is the electrostatic gravel bed filter. In this device, PM in flue gases is removed by impaction
with gravel media inside a packed bed. Collection is augmented by an electrically charged grid within the bed.
Particulate collection efficiencies are typically over 80 percent (USEPA, 2003).
2-21
-------
Table 2-8. Boiler PM Emission Control Technologies
Control Option
Description
Performance
Application
Electrostatic
Precipitator (ESP)
Uses electrical forces to move the
particles out of the gas stream and onto
collector plates. Once the particles are
collected on the plates, they are typically
removed from the plates by knocking,
allowing the collected layer of particles
to slide down into a hopper, which is
later emptied.
-Coal: >99 % PM
reduction
- Oil: 40-60 %
with older ESPs;
up to 90 % with
new or rebuilt
ESPs
-Wood: 90%
PM reduction
ESP technology is applicable to a variety
of coal combustion sources. Because of
their modular design, ESPs can be
applied to a wide range of system sizes
and should have no adverse effect on
combustion system performance.
ESPs are commonly used in oil-fired
power plants and in wood-fired boilers
when collection efficiencies above 90 %
are required.
When applied to wood-fired boilers, ESPs
are often used downstream of
mechanical collector pre-cleaners which
remove larger-sized particles.
Fabric Filter
(or baghouse)
Consists of a number of filtering
elements (bags) along with a bag
cleaning system contained in a main shell
structure incorporating dust hoppers.
-Coal: <99.9%
PM reduction
-	Oik > 99 % PM
reduction
-	Wood: >80 %
PM reduction
Fabric filtration has been widely applied
to coal combustion sources since the
1970s.
Fabric filters have had limited
applications to wood-fired boilers.
Despite complications, fabric filters are
generally preferred for boilers firing salt-
laden wood.
Wet Scrubber
Includes venturi and flooded disc
scrubbers, tray or tower units, turbulent
contact absorbers, and high-pressure
spray impingement scrubbers.
Coal: 95-99 %
PM reduction
Oil: 50-60 % PM
reduction
Wood: >85%
PM reduction
Wet scrubbers are applicable for PM as
well as S02 control on coal- and oil-fired
combustion sources.
The most widely used wet scrubbers for
wood-fired boilers are venturi scrubbers.
Mechanical Collector
Cyclone separators can be installed
singly, in series, or grouped as in a
multicyclone or multiclone collector.
These devices are referred to as
mechanical collectors.
Coal: 90-95 %
PM reduction
Oil: <85 % PM
reduction
Wood: 25-65%
PM reduction
These devices are often used as a pre-
collector upstream of an ESP, fabric filter,
or wet scrubber so that these devices can
be specified for lower particle loadings to
reduce capital and/or operating costs.
For oil combustion, mechanical collectors
are primarily useful in controlling
particulates generated during soot
blowing, during upset conditions, or
when very dirty heavy oil is fired.
Mechanical collectors also provide
particulate control for many wood-fired
boilers. Often, two multiclones are used
in series, allowing the first collector to
remove the bulk of the dust and the
second to remove smaller particles.
Note: data from (USEPA, 1998b, 2002, 2003, and 2010b)
2-22
-------
Fabric filter. Fabric filtration has been widely applied to coal combustion sources since the early 1970s and
consists of a number of filtering elements (bags) along with a bag cleaning system contained in a main shell
structure incorporating dust hoppers. The particulate removal efficiency of fabric filters is dependent on a
variety of particle and operational characteristics. Particle characteristics that affect the collection efficiency
include particle size distribution, particle cohesion characteristics, and particle electrical resistivity. Operational
parameters that affect fabric filter collection efficiency include air-to-cloth ratio, operating pressure loss,
cleaning sequence, interval between cleanings, cleaning method, and cleaning intensity. In addition, the particle
collection efficiency and size distribution can be affected by certain fabric properties (e.g., structure of fabric,
fiber composition, and bag properties). Collection efficiencies of fabric filters can be as high as 99.9 percent for
coal combustion and more than 99 percent for fuel oil combustion (USEPA, 1998b, 2010b).
Fabric filters have had limited application to wood-fired boilers. The principal drawback to fabric filtration, as
perceived by potential users, is a fire danger arising from the collection of combustible carbonaceous fly ash.
Steps can be taken to reduce this hazard, including the installation of a mechanical collector upstream of the
fabric filter to remove large burning particles of fly ash (i.e., "sparklers"). Despite complications, fabric filters are
generally preferred for boilers firing salt-laden wood. This fuel produces fine particulates with a high salt content
having a quenching effect, thereby reducing fire hazards. Particle collection efficiencies are typically 80 percent
or higher (USEPA, 2003).
Wet scrubber. Wet scrubbers, including venturi and flooded disc scrubbers, tray or tower units, turbulent
contact absorbers, or high-pressure spray impingement scrubbers are applicable for PM as well as SO2 control
on coal-fired combustion sources. Scrubber collection efficiency depends on particle size distribution, gas side
pressure drop through the scrubber, and water (or scrubbing liquor) pressure, and can range between 95 and
99 percent for a 2-micron particle (USEPA, 1998b).
Scrubbing systems have also been installed on oil-fired boilers to control both SO2 and PM. These systems can
achieve SO2 removal efficiencies of 90 to 95 percent and particulate control efficiencies of 50 to 60 percent on
oil-fired boilers (USEPA, 2010b).
The most widely used wet scrubbers for wood-fired boilers are venturi scrubbers. With gas-side pressure drops
exceeding 15 inches of water, particulate collection efficiencies of 85 percent or greater have been reported for
venturi scrubbers operating on wood-fired boilers (USEPA, 2003).
Mechanical collector. Cyclone separators can be installed singly, in series, or grouped as in a multicyclone
collector. These devices are referred to as mechanical collectors and are often used as a pre-collector upstream
of an ESP, fa brie filter, or wet scrubber so that these devices can be specified for lower particle loadings to reduce
capital and/or operating costs. The collection efficiency of a mechanical collector depends strongly on the
effective aerodynamic particle diameter. Although these devices will reduce PM emissions from coal combustion,
they are relatively ineffective for collection of particles less than 10 microns (PM10). The typical overall collection
efficiency for mechanical collectors ranges from 90 to 95 percent for coal combustion (USEPA, 1998b).
For oil combustion, mechanical collectors are useful primarily in controlling particulates generated during soot
blowing, during upset conditions, or when dirty heavy oil is fired. For these situations, high-efficiency cyclonic
collectors can achieve up to 85-percent control of particulate. Under normal firing conditions, or when clean oil
is combusted, cyclonic collectors are not nearly so effective because of the high percentage of small particles
(less than 3 micrometers in diameter) emitted (USEPA, 2010b).
2-23
-------
Mechanical collectors also provide particulate control for many wood-fired boilers. Often, two multiclones are
used in series, allowing the first collector to remove the bulk of the dust and the second to remove smaller
particles. The efficiency of this arrangement varies from 25 to 65 percent (USEPA, 2003).
2,3,3,4, Boiler CO2 Reduction
Carbon is a basic component of fossil fuels, not an impurity (like sulfur) or a byproduct of combustion (like
NOx).Removing carbon from flue gases after combustion is therefore energy intensive and extremely expensive.
Thus, for the foreseeable future, there are only two practical ways to reduce CO2 emissions cost effectively from
fossil-fueled combustion: switch to a lower-carbon fuel or increase process energy efficiency so that less fuel is
combusted (STAPPA, 1999). These operational changes often also result in reductions of other air pollutants
such as SO2, NOx, and PM, particularly when GHG emissions are reduced by reducing energy consumption.
Fuel Switching. According to NCASI 2009b, reducingGHG emissions by changingfuels can have significant effects
on SO2, NOx, and PM emissions because the fossil fuels that tend to have the highest sulfur content (coal and
fuel oil) are also the most GHG-intensive. Switching from these fuels to natural gas or biomass would be expected
to reduce SO2 emissions. The effects of fuel selection on NOx emissions are more complex because NOx
emissions are affected not only by fuel type but also by the combustion conditions. Though a significant portion
of the fuel nitrogen can be converted to NOx during combustion, the amount of nitrogen available in the fuel is
relatively small compared with the amount of nitrogen available for conversion in the combustion air. Peak
combustion temperatures influence the magnitude of that conversion (NCASI, 2009b). Some general
information on the GHG intensity relative to the potential for formation of SO2 and NOx emissions associated
with different fuels is presented in Table 2-9.
Table 2-9. Representative Boiler Efficiency and GHG Emission Factors
Fuel
Boiler efficiency (% of fuel
energy [HHV]* transferred
into steam)
C02 Emissions
kg COz/MMBtu
Natural Gas
80
53.02
Distillate Oil (no. 2)
not provided
73.96
Residual Oil (no. 6)
82
75.10
Coal (bituminous)
84
93.4
Bark and Wood Residue
65
93.8a
a Alternatively, zero if credit given for biogenic emissions of C02.
*Higher Heating Value
Reducing GHG emissions by selecting low GHG-intensity fuels can affect PM emissions, with the effect ranging
from strong co-benefits to strong trade-offs. In general, although solid fuels are associated with higher PM
emissions than liquid and gaseous fuels, the emissions are also highly dependent on the type and efficiency of
the device used to control PM emissions. In the US, the two solid fuels used most by the industry, coal and wood-
derived biomass fuels, are at the opposite end of the range of GHG emission factors (assuming that biomass is
treated as carbon neutral). Therefore, fuel switching from coal to biomass, which would greatly reduce GHG
emissions (assuming that biomass is treated as carbon neutral), may not significantly affect PM emissions.
Switching from coal to natural gas would accomplish reductions in both PM emissions and GHGs. At the other
end of the spectrum, switching from natural gas to solid biomass would significantly reduce GHG emissions
2-24
-------
(assuming that biomass is treated as carbon neutral), but in all likelihood, significantly increase PM emissions
(NCASI, 2009b).
Reducing energy-related GHG emissions by fuel switching can affect energy consumption. The amounts of
usable energy obtained from fuels are more or less inversely related to the GHG emissions of the fuel, as shown
in Table 2-9. For example, a change from coal to bark will accomplish a very large reduction in GHG emissions
(assuming biomass is considered carbon neutral) but will require more total energy consumption because more
bark is required to produce the same amount of usable energy (due to its high moisture content). Switching to
less GHG-intensive fuels seldom reduces total energy consumption, although it can significantly reduce non-
renewable energy consumption if the change involves switching from a fossil fuel to biomass.
A 2001 report by NCASI describes the applicability and limitations of switching a power boiler from fossil fuel to
wood fuel, or to build a new boiler to utilize available biomass fuel. Fuel switching of fossil fuel-fired power
boilers to biomass would reduce on-site CO2 generation, given that biomass fuels are considered to have a net
zero CO2 emission factor. Separately, the NCASI report presents details on switching a power boilerfrom coal or
oil to natural gas (NCASI, 2001).
Energy Efficiency Measures. Numerous energy efficiency measures may be applicable for steam and power
supply systems in pulp and paper plants (Andover, 2010; NCASI, 2001; USEPA, 2010). Some measures such as
the ones given below require capital investment.
•	Replacing low pressure boilers and installing turbogenerator capacity
•	Replacing burners
•	Preheating demineralized water with secondary heat before steam heating
•	Rebuilding or replacing low efficiency boilers
•	Installing a steam accumulator to facilitate efficient control of steam header pressures
•	Installing an ash reinjection system in the hog fuel boiler
•	Installing a bark press or bark dryer to increase utilization of biofuels
•	Installing additional heat recovery systems on boilers to lower losses with flue gases
•	Installing a gas turbine cogeneration system for electrical power and steam generation
•	Installing flue gas heat recovery systems
•	Improving boiler insulation
•	Implementing condensate return
Other measures mostly require improved operation and maintenance practices such as:
•	Implementing energy management program for current and reliable information on energy use
•	Improved boiler and process control and maintenance
•	Steam trap maintenance and automatic steam trap monitoring
2-25
-------
•	Leak repair (LR)
•	Minimizing boiler blow down, and blow down steam recovery
•	Reducing of excess air amount and flue gas quantities
The above information could be reviewed further to determine appropriate energy efficiency measures for
inclusion in the Universal ISIS-PNP for purposes of assessing CO2 emission reductions and their associated costs.
2.4. Recovery Furnaces
2.4.1.
Chemical recovery is the heart of the kraft mill that allows the kraft mill to operate as an essentially closed
operation with recovery of spent cooking chemicals to produce fresh cooking liquor (Na2S and NaOH) (NCASI,
2009a). See Figure 2-5 for an illustration of the chemical recovery process. Further information about the
process is provided below, based on a background document developed by USEPA in support of the chemical
recovery combustion sources NESFIAP (USEPA, 1996).
The Sodium Loop	The Calcium Loop
Bionias s
CaO
Fossil and
Biomass
Lime Mud
sCaC03
CaO+H2G Ca(OH)2
Slaker
CaCG3-*CaQ4-CG2T
Fossil fuel + 02 C02t
Lime Kiln
Causticizers + White Liquor Clarifier
Na2CG3 + Ca(OH)2 2NaOH + CaC03-l
Recovery Furnace + Smelt
Di ssolving Tank
Wood Orgaiiics + 02 C02
Na end S Cpds.-i* Na2S
Na Cpds + C02-» Na2C03
Pulping Digester
NaOH + Na2S + wood chips
Various Na and S Cpds, Pulp fibers,
and Dissolved Wood Material.
Figure 2-5. Simplified Representation of the Kraft Pulping and Chemical Recovery System (USEPA,2010a)
In the chemical recovery process, weak black liquor from pulp washing is first directed through a series of
multiple-effect evaporators to increase the solids content to approximately 50 percent. The strong black liquor
from the multiple-effect evaporators system is then either oxidized in the black liquor oxidation system, it is
further concentrated in a direct contact evaporator, or routed to a non-direct contact evaporator, also called a
concentrator. Oxidation of the black liquor in the black liquor oxidation system stabilizes the sulfur compounds
in the black liquor by converting the Na2S in the liquor to nonvolatile sodium thiosulfate, thereby reducing
emissions of TRS compounds, which are stripped from the black liquor in the direct contact evaporator when it
contacts hot flue gases from the recovery furnace. Black liquor that is concentrated in non-direct contact
evaporators does not contact the hot flue gases and, therefore, does not require oxidation. The solids content
of the black liquorfollowing the final evaporator/concentrator is 65 percent or higher (USEPA, 1996).
2-26
-------
The concentrated black liquor is then fired into the recovery furnace (sometimes referred to as a "recovery
boiler"), where organic compounds are combusted, and the sodium sulfate (Na2S04) in the black liquor is
reduced to Na2S. Since the 1970s, most new recovery furnaces have been designed with non-direct contact
evaporators (USEPA, 1996). Over 70 percent of recovery furnaces currently in operation are non-direct contact
evaporator furnaces, and less than 30 percent are direct contact evaporator furnaces, based on the latest
inventory information (RTI, 2013a).
The black liquor burned in the recovery furnace has a high energy content, which is recovered as steam for
process requirements such as cooking wood chips, heating and evaporating black liquor, preheating combustion
air, and drying the pulp or paper products. When necessary, natural gas or distillate oil is used as an auxiliary
fuel (usually for furnace startup and shutdown). Particulate matter (primarily Na2S04 and sodium carbonate
[Na2CC>3]) exiting the recovery furnace with the hot flue gases is collected in an ESP and added to the black liquor
to be fired in the furnace. Additional makeup Na2S04, or "salt cake," may also be added to the black liquor prior
to firing. Molten inorganic salts, referred to as "smelt," are one of the main products from the combustion of
black liquor, and they collect in a char bed at the bottom of the recovery furnace. Smelt is drawn off and
dissolved in weak wash water in the smelt dissolving tank associated with the furnace to form a solution of
carbonate salts called "green liquor," which is primarily Na2S and Na2C03. Reprocessing of the green liquor into
cooking liquor continues after the smelt dissolving tank (USEPA, 1996).
2,4,2, S ' 2 " • . • • . " ' "
A background discussion of the NOx, SO2, and PM emissions from kraft recovery furnaces is provided in the
paragraphs below, based on information from two NCASI reports (NCASI, 2004, 2009a).
The basic elements of pulping chemicals are sulfur and sodium. The recovery furnace is designed and must be
operated to maximize capture of these substances, as well as separate and burn the organic substances
dissolved from wood chips during pulping. The chemistry progresses through a series of complex reactions
responsive to temperatures and the staged addition of combustion air that regulates available oxygen levels
over the height of the furnace. The furnace environment is non-uniform (NCASI, 2009a).
Temperatures and oxygen-deficient reducing conditions at the base of the furnace produce molten Na2S.
Sodium fumes released in that region of the furnace react with SO2 formed higher in the furnace, where excess
oxygen levels are conducive to oxidation of hydrogen sulfide (H2S) that also originates in the furnace reducing
zone. Emissions of sulfur are related to the composition of the spent pulping liquor being recovered and the
staged combustion conditions in the furnace. Nitrogen compounds will also be liberated from the liquor in the
lower furnace and, depending upon temperatures, may take a form that contributes to greater formation of
NOx in the furnace. Except for very limited circumstances, recovery furnace temperatures do not reach levels
that support the oxidation of combustion air nitrogen to form NOx. Thus, emissions of NOx are related to the
composition of the spent pulping liquor being recovered and the staged combustion conditions in the furnace
(NCASI, 2009a).
The above description of recovery furnace chemical reactions illustrates circumstances that contribute to
emissions of NOx and SO2. A host of other chemical reactions occur as combustion gases rise through the various
zones of the furnace. The conditions under which these reactions occur influence emissions of not only NOx, but
also odorous reduced sulfur gases, carbon monoxide (CO), VOCs, and other compounds of environmental
interest. The emission levels of these various substances are inter-related and cannot all be simultaneously
2-27
-------
controlled to low levels by manipulation of combustion conditions. Nor can sight be lost of the purpose of the
furnace in the closed-loop recovery of pulping chemicals (NCASI, 2009a).
2.4.2.1.	Recovery Furnace NOx Emissions
Recovery furnace NOx emissions are influenced by pulping liquor nitrogen content, combustion temperatures in
the reducing zone of the furnace, and excess oxygen in the zone where most of the liquor combustion occurs.
Recovery furnace SO2 emissions are a function of liquor properties such as sulfidity (sulfur-to-sodium ratio),
solids content and associated heat value; combustion air and liquor firing patterns; furnace design features;
furnace load; auxiliary fuel use; and stack gas oxygen content. None of these factors, however, exhibit a
consistent relationship with SO2 emissions (NCASI, 2009a).
Kraft recovery furnaces typically have inherently low NOx emissions due to the following factors (NCASI, 2009a):
•	Low nitrogen (N2) concentrations in most "as-fired" black liquor solids (< 0.2 %)
•	Low overall conversions of liquor N2 to NOx by the fuel NOx formation pathway
•	Insufficient temperatures for thermal NOx formation, and perhaps
•	Highly staged combustion design of recovery furnaces
•	Existence of sodium fumes that might promote "in-furnace" NOx reduction or removal.
Overall conversions of black liquor nitrogen to nitric oxide (NO) are quite low compared with otherfuels, ranging
from 10 to approximately 25 percent. Emission levels for individual furnaces do not vary greatly. However, there
can be wide differences from one furnace to another, reinforcing the observation that each recovery furnace is
an individual and that optimum conditions for process and emission performance must be carefully sought
(NCASI, 2009a).
2.4.2.2.	Recovery Furnace SO2 Emissions
Conditions involving liquor quality (such as high Btu, high solids content, and sulfidity), liquorfiring patterns, and
conditions related to furnace operations (air distribution, auxiliary fuel, etc.) that lead to maximizing
temperatures in the lower furnace generally result in minimizing SO2 emissions from kraft recovery furnaces.
Emissions of SO2 are typically less than 100 parts per million (ppm) and are extremely variable, a measure of the
dynamic nature of furnace operations (NCASI, 2009a).
In general, recovery furnaces are not used to incinerate waste streams generated in other parts of the mill, with
the exception of some recovery furnaces that receive high-volume, low-concentration NCG containing TRS from
the pulping area of the mill. Because of the importance of the recovery furnace to the chemical recovery cycle
and the potential for catastrophic explosion (due to water entering the furnace during operation), the industry's
Black Liquor Recovery Boiler Advisory Committee has recommended that the recovery furnace not be used for
NCG incineration (NCASI, 2009a). Thus, the contribution of TRS-containing NCG streams to SO2 formation is not
a consideration for most recovery boilers.
2.4.2.3.	Recovery Furnace PM Emissions
Recovery furnaces are designed and operated to ensure the presence of high levels of sodium fumes to capture
the SO2 produced as a result of oxidation of reduced sulfur compounds. Consequently, recovery furnace flue
gases contain high levels of PM. The uncontrolled PM load from recovery furnaces is highly variable and has
2-28
-------
been reported to range from 100 to 250 pounds per oven-dried ton of pulp for direct contact evaporator
furnaces and 200 to 450 pounds per oven-dried ton of pulp for non-direct contact evaporator furnaces. The
lower particulate loading from direct contact evaporator furnaces is due to the capture of some PM in it (NCASI,
2004). However, use of direct contact evaporator recovery furnaces is being phased out in favor of more energy
efficient non-direct contact evaporator systems that produce fewer emissions of pollutants such as TRS, VOCs,
and HAPs.
Particulates generated in the recovery furnace are comprised mainly of Na2S04, with lesser amounts of Na2CC>3
and sodium chloride (NaCI). Similar potassium compounds are also generated, but in much lower a mounts. Trace
amounts of other metal compounds, e.g., magnesium, calcium, and zinc, can be present. A significant portion of
the particulate material is sub-micron in size, which makes removal with add-on control devices more difficult
(NCASI, 2004).
Increasing liquorfiring density (ton/day/square foot) has been reported to increase recovery furnace particulate
loading. Other factors such as bed and furnace temperature, liquor solids, liquor composition, and air
distribution also affect uncontrolled particulate emissions from recovery furnaces (NCASI, 2004).
2,4,2,4, Recovery Furnace GHG Emissions
Concentrated spent pulping liquors generated as a byproduct of chemical pulping are burned in chemical
recovery furnaces (or other types of chemical recovery combustion units) to produce steam for use in facility
processes and to recover chemicals for reuse in the pulping process. Carbon dioxide emissions associated with
combustion of spent pulping liquor (e.g., black liquor at kraft mills) in chemical recovery furnaces are biomass-
derived CO2 because the carbon originates from the wood or other cellulosic materials. The carbon in the spent
pulping liquor exits the recovery furnace in two forms: (1) as CO2 emissions from the recovery furnace stack, and
(2) as carbonates in the smelt flowing from the bottom of the recovery furnace (which eventually makes its way
to the lime kiln) (USEPA, 2009).
Small amounts of supplemental fossil fuels (e.g., oil or natural gas) are also fired in the furnace, usually during
startup or shutdown conditions. Therefore, chemical recovery furnaces are sources of both biogenic and fossil
fuel-based CO2 as well as small amounts of CH4 and N2O. National statistics indicated that 98 percent of the
annual heat input to chemical recovery furnaces originated from biomass in 2005 (USEPA, 2009). Thus, Universal
ISIS-PNP could focus on biogenic CO2 emissions from the recovery furnace (from combustion of spent liquor) or
exclude recovery furnace CO2 reduction measures from consideration until future Universal ISIS versions are
developed to utilize an accounting methodology that does not consider biogenic CO2 emissions to be zero.
2.43, Recovery Furnace Emission Reduction Strategies
The subsections below discuss measures for reducing emissions of NOx, SO2, PM, and GHG (predominantly CO2)
from recovery furnaces. Distinctions in recovery furnace type (direct versus non-direct contact evaporator) are
made where appropriate. As was mentioned previously, that non-direct contact evaporator recovery furnaces
do not appear to significantly reduce PM, SO2 or NOx emissions, but that they are preferred over direct contact
evaporators because of their impact on TRS (STAPP, 2006).
2,4,3,1, Recovery Furnace NOx Reduction
According to NCASI 2009a, optimization of staged combustion within a large existing kraft recovery furnace to
obtain from 20-to 40-percent reduction in prevailing NOx emissions is the only technologically feasible reduction
measure that has been demonstrated at the present time. However, the report cautioned that the effects of
2-29
-------
such air staging on emissions of other pollutants, mainly TRS, SO2, and CO, and on other furnace operational
characteristics, including fouling, plugging, and chloride buildup, need to be examined with longer-term data. In
addition, lowerfurnace temperature conditions conducive to low NOxformation aggravate SO2 emissions (NCASI,
2009a). The report concluded that many of the commonly cited NOx control options (such as low NOx burners
or SNCR) can be dismissed either because they are inappropriate for the nature of recovery furnace NOx
formation or incompatible with recovery furnace chemistry and operational integrity (NCASI, 2009a). SNCR and
SCR technologies for recovery furnaces have been investigated but have been determined not to be technically
feasible. Low-NOx burners appear to affect efficiency and energy usage adversely; staged combustion has been
determined to be best available control technology (BACT) in at least one state (STAPPA, 2006). Attributes of
potential recovery furnace NOx control technology options are summarized in Table 2-10.
2-30
-------
Table 2-10. Recovery Furnace NOx Control Technologies
Control Option
Description
Performance
Application
Combustion Modifications
Overfire Air
(OFA)
Optimizing staged combustion in the upper furnace
reduces availability of oxygen for oxidation of nitrogen
compounds originating in the pulping liquor. Limited
short-term experience after installing "quaternary" air
ports (overfire air) in two US furnaces showed that a
20-40 % reduction in baseline NOx levels is feasible.
Comparable performance has been reported abroad.
The practice would be limited to large furnaces. The
reduction of NOx emissions is variable, dependent on
the furnace type and design and the method of OFA
application. OFA has to be adapted to the specific
conditions of recovery furnaces.
20-40 % NOx
reduction
The application of this technique may
result in increases in CO and unburned
carbon emissions if not well controlled.
The effect of such air staging on
emissions of other pollutants, chiefly
SO2, CO, and TRS, and other furnace
operational characteristics needs to be
examined with longer-term data on
North American furnaces.
Direct contact evaporator recovery
furnaces are smaller and may not have
room for additional levels of air (USEPA,
2012).
Low-NOx Burners
(LNB)
The highly staged combustion design of recovery
furnaces, the inherent low reducing zone oxygen
concentrations needed for efficient recovery of
chemicals, and the dominance of temperature-
sensitive fuel nitrogen precursors of NOx combine to
render low-NOx burners unproductive.
Infeasible

Oxygen Trim and
Water Injection
Neither option is appropriate for kraft recovery
furnaces since: (a) any injection of water into the
furnace would lead to an unacceptable explosive
condition; and (b) the oxygen trim technique would
have marginal effect due to the already existing highly
staged combustion air configuration in recovery
furnaces.
Infeasible

Flue Gas
Recirculation
(FGR)
In FGR, a portion of the uncontrolled flue gases is
routed back to the combustion zone, primarily with
the intention of reducing thermal NOx. Recovery
furnace NOx emissions are dominated by nitrogen
that originates in the black liquor, not the oxidation of
nitrogen in combustion air. Operational handicaps and
other means for reducing fuel-related NOx erode the
viability of FGR on recovery furnaces.
Infeasible
FGR would add additional gas volume
in the furnace, increasing velocities and
potentially causing more liquor
carryover, which would result in
increased fouling of the recovery
furnace tubes.
Post-Combustion/Flue Gas Treatments
Selective
Catalytic
Reduction (SCR)
The use of SCR on a kraft recovery furnace has never
been demonstrated, even on a short-term basis. The
impact on catalyst effectiveness of high PM
concentrations in the economizer region of the
furnace and fine dust particles is a major impediment
to the application of this technology ahead of PM
control. Installation after the PM control device would
render the gas stream too cold for effective reaction
with the NOx. Catalyst poisoning by soluble alkali
metals in the gas stream is also problematic (NCASI
2009a).
Consequently, it would be necessary to install the SCR
after removal of catalyst poisons from the gas stream,
Traditional SCR:
infeasible (STAPPA,
2006)
Tail-end SCR:
unattractive due to
additional fuel
requirements
(Andover, 2010)
RSCR: recently
developed, >75 %
Reheating the flue gas after the
particulate control device and ahead of
the SCR section would incur a
substantial energy penalty, which could
be reduced by using a RSCR to more
efficiently reheat the gas (Andover,
2010).
2-31
-------
Control Option
Description
Performance
Application

suggesting that a tail-end SCR would be needed.
However, tail-end SCRs are unattractive due to the
additional fuel necessary for reheating the gas. An
alternative to a traditional "tail end" SCR is a RSCR,
which applies SCR in combination with regenerative
thermal oxidizer technology to reheat cleaned gas to
SCR operating temperatures more efficiently. RSCR is
a recently-developed technology that has been used
on biomass-fired boilers downstream of PM removal
devices to reduce NOx by over 75 % (Andover 2010).
NOx reduction for
biomass boilers
(Andover, 2010),
but not currently
demonstrated for
recovery furnaces

Selective Non-
Catalytic
Reduction (SNCR)
SNCR, which uses the injection of urea or ammonia
into a high temperature location in the furnace, is not
considered technologically feasible for recovery boiler
applications because of the risk of disrupting the
complex chemistry of the unit. Trials with ammonia
injection in Europe indicate a 30 % NOx removal
capability.
Infeasible (STAPPA,
2006)
Because the use of urea can eventually
cause corrosion problems due to the
possible formation of corrosive
byproducts, safety concerns
discourage, if not preclude, its use in
recovery boilers.
Low-
Temperature
Oxidation
(LoTOx)
LoTOx is a process whereby the NOx compounds are
oxidized to water-soluble forms, which are
subsequently captured in a downstream wet
scrubber. Can achieve higher than 90 % NOx reduction
(Andover, 2010).
Potentially >90 %
NOx reduction
(Andover, 2010),
but not currently
demonstrated for
recovery furnaces
This process can only perform in
combination with a downstream wet
scrubber (Andover, 2010).
Note: all data from NCASI 2009a, except where otherwise indicated.
2,4.3,2, Recovery Furnace SO2 Reduction
According to NCASI 2009a, firing more concentrated black liquor is conducive to reduced SO2 emissions, but the
report cautioned that this increases NOx formation and particulate emissions, requiring additional control. The
report concluded that alkaline scrubbing is the most viable post-combustion control option, with reported
removals up to 90 percent (NCASI, 2009a). However, the report cautioned that conducting alkaline scrubbing
with the many furnaces that emit low levels of SO2 (20 ppm and less) would be very difficult and extremely
expensive due to the large gas volumes involved. The report concluded that scrubbing would not be a realistic
alternative for those recovery furnaces, nor would 90-percent reductions be achievable in that circumstance.
The report noted that flue gas treatment for SO2 reduction has been applied abroad but not in North America
(NCASI, 2009a).
NCASI 2009a noted that potential dividends associated with alkaline scrubbing include increased retention of
process sulfur and heat recovery, in cases where it can be used. However, the report indicated that any
associated capture of process sulfur and heat, as well as avoidance of a wastewater stream, would depend upon
the available capacity of equipment components associated with the pulping liquor recovery process (NCASI,
2009a).
Attributes of potential recovery furnace SO2 control technology options are summarized in Table 2-11.
2-32
-------
Table 2-11. Recovery Furnace S02 Control Technologies
Control Option
Description
Performance
Application
Increasing Black
Liquor
Concentration
Maximizing temperatures in the lower
furnace by combustion of more
concentrated liquor enhances the
formation of Na2S04, with a concurrent
gaseous S02 reduction.
S02 reduction
site-specific
Increased lower furnace temperatures
associated with more concentrated liquor
firing increase conversion of fuel nitrogen
to NO. That phenomenon, combined with
a possible greater tendency for the
creation of thermal NOx and diminished
capability for internal alkaline fume
capture of NOx, results in greater furnace
NOx emissions. Increasing black liquor
dissolved solids content from a common
65 % up to 75 % may increase NOx
emissions by up to 20 %.
Firing more concentrated liquor increases
the emissions of particulates prior to flue
gas cleaning. To compensate, a more
efficient and expensive electrostatic
precipitator has to be installed.
Concentrating solids may liberate sulfur
compounds, requiring collection and
incineration, producing S02.
Scrubber
The few scrubbers that exist on recovery
furnaces in the US pulp and paper
industry were installed for purposes
other than SO2 control and do not reflect
the range of capability. Experience
abroad indicates removal efficiency for
S02 in excess of 90 %.
Greater than 90
% S02 reduction
(outside US); 90
% reduction not
expected to be
achievable for
furnaces with low
S02 levels
The scrubber requires alkali in the form of
oxidized white liquor, weak liquor or
NaOH, which can increase the capacity
demands on other components of the
chemical recovery process.
Note: all data from NCASI 2009a.
2,4,3,3, Recovery Furnace PM Reduction
The following paragraphs summarize available information on PM control technologies for kraft recovery
furnaces from USEPA's background document for the chemical recovery combustion sources NESHAP (USEPA,
1996).
Due to State and Federal regulations regarding PM emissions and the economic benefits of recycling PM
captured from the recovery furnace flue gases, all recovery furnaces are equipped with add-on PM control
devices as baseline controls. Electrostatic precipitators are a demonstrated control technique for reducing PM
emissions from recovery furnaces. Particulate matter emissions from approximately 96 percent of all recovery
furnaces are controlled with an ESP alone. The remaining furnaces are controlled with ESPs followed by wet
scrubbers (3 %) or with wet scrubbers alone (USEPA, 1996; RTI, 2013a).
Properly designed and operated ESPs used on recovery furnaces routinely achieve PM removal efficiencies of
99 percent or greater. The ESPs used to control PM emissions from recovery furnaces are generally classified as
plate-wire ESPs where the flue gas flows between parallel sheet metal plates and high-voltage electrodes. Each
paired set of electrodes and plates forms a separate electrostatic field within the ESP. The ESPs used to control
2-33
-------
recovery furnace PM emissions typically have two parallel precipitator chambers (i.e., flue gas passages) with
three or four electrostatic fields per chamber (USEPA, 1996).
The PM recovered in the ESP (Na2S04, with lesser amounts of Na2CC>3 and NaCI) is subsequently recycled to the
black liquor. The ESPs used on recovery furnaces may be designed with either a wet or dry bottom. In wet-
bottom ESPs, the collected PM falls directly into a pool of liquid, which may be black liquor or water, in the
bottom of the ESP. In dry-bottom ESPs, the collected PM falls to the (dry) bottom of the ESP and is transferred
from the ESP bottom to a mix tank (containing black liquor) via drag-chain or screw conveyors. Black liquor is
sometimes used to transport the dry collected PM to the mix tank. More recent ESP installations employ a dry
PM return system to transport the PM to the mix tank. Because the PM removed by the ESP is recycled to the
black liquor in the mix tank, the ESP is an integral part of the chemical recovery loop as well as an air pollution
control device (USEPA, 1996).
The two recovery furnace types often differ in the types of ESPs that are used to control PM emissions from the
furnace. For example, the ESPs that control PM emissions from direct contact evaporator recovery furnaces tend
to be wet-bottom ESPs, whereas ESPs on non-direct contact evaporator recovery furnaces tend to be dry-
bottom ESPs, with wet or dry PM return systems (USEPA, 1996).
The average lifetime of an ESP in service on a recovery furnace varies depending upon the type of ESP bottom
(i.e., wet vs. dry), the inlet temperature of the gases, and maintenance practices. The lifetime of ESPs used to
control PM emissions from recovery furnaces with non-direct contact evaporators, which tend to operate with
dry-bottom ESPs, typically ranges from 12 to 15 years. After that point, major repairs or a rebuild may be
required. Recovery furnaces with direct contact evaporators tend to have cooler inlet gases and wet-bottom
ESPs; these two factors promote corrosion through condensation of acid gases and shorten the life of the ESP
to from 7 to 10 years (USEPA, 1996).
Opportunities may exist to reduce current PM emissions levels by upgrading or replacing older ESPs or adding a
wet scrubber after the ESP. For example, a STAPPA 2006 document noted that, on recovery furnaces, older
model ESPs have collection efficiencies close to 90 percent, while newer model ESPs have collection efficiencies
greater than 99 percent. The STAPPA 2006 document further stated that the cost of retrofitting a recovery
furnace with an ESP is heavily influenced by site-specific factors (STAPPA, 2006).
A few recovery furnaces currently employ both an ESP and wet scrubber, which can simultaneously reduce
emissions of sulfur compounds (SO2, TRS) and PM. A review of available PM emissions data for these recovery
furnaces suggests that wet scrubbing of recovery furnace exhaust gases (either alone or in conjunction with an
ESP) does not necessarily improve filterable PM removal. The wet scrubbers installed following recovery furnace
ESPs are typically designed for SO2 removal rather than for removal of PM (RTI, 2013a).
Attributes of potential recovery furnace PM control technology options are summarized in Table 2-12.
2-34
-------
Table 2-12. Recovery Furnace PM Control Technologies
Control Option
Description
Performance
Application
ESP Upgrade/
Replacement
Upgrading existing ESP
or replacing older ESP
with new ESP
Could increase PM control from 90
to 99 % (STAPPA 2006)
The cost of retrofitting a recovery furnace
with an ESP is heavily influenced by site-
specific factors including the age and design
of the ESP already in place.
Addition of Wet
Scrubber
Adding a new wet
scrubber after an
existing recovery
furnace ESP
No improvement in PM control
found. A review of the available PM
data showed an overlap in PM
control efficiencies and PM
emissions for recovery furnaces
equipped with an ESP and those
equipped with an ESP + wet
scrubber. Also, the best-performing
ESP-wet scrubber (99 %) is less
efficient than the best-performing
ESP alone (99.96 %) (RTI 2013b).
Some recovery furnaces equipped with a
wet scrubber alone or with a wet scrubber
in combination with an ESP exhibited PM
emissions above the NESHAP new source
PM limit of 0.015 grains per dry standard
cubic foot (gr/dscf), but below the NESHAP
existing source PM limit of 0.044 gr/dscf,
suggesting that wet scrubbing of recovery
furnace exhaust gases (either alone or in
conjunction with an ESP) does not
necessarily improve filterable PM removal.
The wet scrubbers installed following
recovery furnace ESPs are typically designed
for SO2 removal rather than for removal of
PM (RTI 2013a).
Note: The PM emissions and control efficiencies for best-performing ESP and ESP-wet scrubber from (RTI 2013b).
2,4,3,4, Recovery Furnace CO2 Reduction
Unlike the situation in boilers, fuel switching is generally not an option of significance for recovery furnaces
because spent pulping liquor comprises most of the heat input. Only small amounts of supplemental fossil fuels
(e.g., oil or natural gas) are fired in the furnace. Energy efficiency measures for recovery furnaces are described
below (USEPA, 2010a).
Black liquor solids concentration. Black liquor concentrators are designed to increase the solids content of black
liquor prior to combustion in a recovery furnace. Increased solids content means less water must be evaporated
in the recovery furnace, which can increase the efficiency of steam generation substantially. Two primary types
of black liquor concentrators are in use today: submerged tube concentrators and falling film concentrators.
In a submerged tube concentrator, black liquor is circulated in submerged tubes, where it is heated but not
evaporated; the liquor is then flashed to the concentrator vapor space, causing evaporation. One study
estimated that, for a 1,000 ton per day pulp mill, increasing the solids content in the black liquor from 66 to 80
percent would lead to fuel savings of 30 MMBtu/h, or approximately $550,000. Capital costs of the high solids
concentrator would include concentrator bodies, piping for liquor and steam supplies, and pumps.
A tube-type falling film evaporator operates almost exactly the same way as a more traditional rising film
evaporator, except that the black liquor flow is reversed. The falling film evaporator is more resistant to fouling
because the liquor is flowing faster, and the bubbles flow in the opposite direction of the liquor. This resistance
to fouling allows the evaporator to produce black liquor with considerably higher solids content (up to 70 %
solids, rather than the traditional 50 %), thus eliminating the need for a final concentrator. One study estimated
a steam savings of 0.76 MMBtu per ton of pulp with this type of concentrator.
2-35
-------
According to another study, a 900-ton-per day-pulp and paper mill that installed a liquor concentrator increased
its solids content from 73 to 80 percent and reduced annual energy usage by about 110,000 MMBtu. Cost
savings for the mill were approximately $900,000 per year, with an estimated payback period of four years
(USEPA, 2010a).
Additional information on the costs and CO2 reductions associated with installing a high solids concentrator to
maximize steam generation with black liquor is provided in NCASI 2001. This report notes that high-solids firing
is applicable only for the non-direct contact evaporator recovery furnaces. The report also describes the impact
of converting a recovery furnace from direct to a non-direct contact evaporator furnace along with
implementation of high solids firing (NCASI, 2001).
Improved composite tubes for recovery furnaces. Recovery furnaces consist of tubes that circulate pressurized
water to permit steam generation. These tubes are normally made of carbon steel, but severe corrosion thinning
and occasional tube failure has led to the research and development of more advanced tube alloys, including
new weld overlay and co-extruded tubing alloys. Replacing carbon steel tubes in the recovery furnace with these
composite alloy tubes allows the use of black liquor with higher dry solids content, which increases the thermal
efficiency of the recovery furnace and decreases the number of furnace shutdowns. Improved composite tubes
have been installed in more than 18 kraft recovery furnaces in the US, leading to a cumulative energy savings of
4.6 trillion Btu since their commercialization in 1996 (USEPA, 2010a).
Recovery furnace deposition monitoring. Better control of deposits on heat transfer surfaces in recovery
furnaces can lead to higher operating efficiency, reduced downtime (by avoiding plugging), and more predictable
shutdown schedules. A handheld infrared inspection system is currently available that can provide early
detection of defective fixtures (tube leaks or damaged soot blower) and slag formation, preventing impact
damage and enabling cleaning before deposits harden. The system can reportedly provide clear images in highly
particle-laden boiler interiors and enable inspection anywhere in the combustion chamber. As of 2005, 69 units
were in use in the US, generating 1.4 tril I ion Btu in energy savings since their introduction in 2002 (energy savings
are attributable to reduced soot blower steam use) (USEPA, 2010a).
Quaternary air injection. Most recovery furnaces in the US have three stages of air injection but use the third
stage in a limited fashion. Using the third stage fully and adding a fourth air injection port can reduce carryover
and tube fouling, thereby reducing the frequency of recovery furnace washing, which will lead to energy savings
because boiler shutdowns and reheating can be reduced. One estimate indicated each boiler reheat cycle will
consume approximately 10 MMBtu at a cost of approximately $50,000. Capital costs for this measure are
estimated at $300,000 to $500,000 (USEPA, 2010a).
2.5. Lime Kilns
2.5.1. [
A background discussion of lime kilns is provided in the paragraphs below based on information from NCASI
2009a. The report noted that the smelt that flows from the recovery furnace consists principally of Na2S and
Na2CC>3. The smelt is combined with wash water to form an intermediate solution known as green liquor, which
requires further chemical processing (or recausticizing) to regenerate pulping liquor. Recausticizing involves the
slaking of quicklime (CaO) into the green liquorto form a solution of Na2S and NaOH known as white liquor. The
2-36
-------
chemical reaction leaves a suspension of calcium carbonate (CaCOa) that is subsequently separated from the
white liquor to complete the liquor recovery cycle (NCASI, 2009a). See Figure 2-5, above, for details.
The separated CaCC>3, known as lime mud, is washed and filtered and reprocessed through the lime kiln to form
CaO to be used again in the recausticizing cycle. The conversion to quicklime involves the burning of lime mud
most often in a rotary lime kiln. Fuels most commonly employed for lime kilns include oil and natural gas, which
may be supplemented by other fuels such as petroleum coke, a carbonaceous byproduct of the oil refining
coking process, or tire-derived fuel.
Rotary lime kilns are large refractory-lined steel cylinders that are slightly inclined from the horizontal and are
slowly rotated. Lime mud is introduced at the higher end and slowly makes its way to the lower discharge end
due to the inclination and rotation. Lime mud and combustion gases flow in opposite directions. The burner is
installed at the discharge end of the kiln. Fleat transfer from this flame and the hot combustion gases that flow
up the kiln dry, heat, and calcine the counter-flowing lime solids (NCASI, 2009a).
In the kiln, the temperature profile from the inlet to the outlet is an important variable that must be controlled
properly to ensure consistent lime quality and reduce operational problems in reaction chemistry. Solids
temperatures range from 80 °C in the drying zone at the feed inlet end of the kiln to higher than 870 °C in the
calcining zone toward the outlet end of the kiln. Primary airflow, apartfrom supporting combustion, is important
for effective heat transfer in the kiln (NCASI, 2009a).
2.5,2. 5
The following paragraphs discuss the sources of lime kiln NOx, SO2, and PM emissions, based on information
from NCASI 2009a and other sources.
Emissions of NOx and SO2 from lime kilns are relatively low. The NOx and SO2 emissions are influenced by fuel
choice, the composition of materials fed to the kiln, chemical reactions that accompany lime mud calcination,
and choice of external control approaches for PM emissions. The report stated that combustion process
modifications may be useful, but cautioned that they are limited by site-specific considerations and product
quality impact (NCASI, 2009a).
2,5,2,1. Lime Kiln NOx Emissions
Though the mechanisms differ, NOx produced in the kraft lime kiln originates from the combustion of fossil fuels.
The formation of NOx is related to the nitrogen content of the fuel. Burner design and flame temperature are
also prominent factors in NOx emissions due to the need to attain a high flame temperature for good heat
radiation to the bed of lime (NCASI, 2009a).
Natural gas and fuel oil, either alone or in combination, are the most common fuels currently used in lime kilns
(RTI 2013a). According to NCASI 2009a, the range of NOx emissions is wide, and data are unclear as to whether
gas or oil is associated with the greater level. The introduction of other fuels (e.g., solid petroleum coke) and
process streams bearing reduced sulfur compounds such as stripper off-gases, which are relatively rich in
nitrogen content, increases the potential for NOx emissions (NCASI, 2009a).
Petroleum coke has between 1.0- and 2.6-percent nitrogen (N2) compared with approximately 0.1- 1.0-percent
N2 for residual fuel oil. The nitrogen content of natural gas is considered insignificant. Thus, there would appear
to be significant potential for fuel NOx formation from petroleum coke combustion. Flowever, observed levels
of NOx emissions from burning petroleum coke in lime kilns suggest that less than 10 percent of the N2 in
2-37
-------
petroleum coke converts to NOx, a level even lower than typical fuel nitrogen conversions for residual fuel oil.
Thus, the report concludes that firing petroleum coke contributes little, if any, increase in NOx emissions (NCASI,
2009a).
Lime kiln SO2 is formed from the combustion of fuel oil, residual sulfide in the lime mud, or other gaseous
streams (NCGs, stripper off-gases) that may be burned in the kiln for purposes of TRS emissions control.
According to NCASI 2009a, kiln chemistry provides a built-in mechanism for SO2 control. The report noted that
sodium liberated from the residual Na2CC>3 in the lime mud combines with SO2 to form Na2SC>4 that is captured
in the kiln particulate control device or retained by the solids in the kiln. The report cautioned that the potential
of this mechanism is not unbounded, however; SO2 reduction will cease once the Na2CC>3 capacity of the mud is
exhausted. Moreover, if the lime mud contains excessive sodium, impaired kiln operation can occur due to
severe ring formation that obstructs kiln operation. Ring formation is a consequence of Na2SC>4 formation in the
kiln lime bed. The control of kiln particulate emissions by wet scrubbers can contribute additional SO2 control
attributable to the alkaline nature of the particulate catch. Consequently, lime kiln SO2 emissions, on average,
are very low (about 50 ppm) (NCASI, 2009a).
2.5.2.2,	Lime Kiln SO2 Emissions
According to NCASI 2009a, the impact of petroleum coke burning on SO2 emissions from lime kilns can be
insignificant in spite of the relatively high levels of sulfur (S) in petroleum coke, 4.9 percent on average. As with
other kiln sulfur inputs, the report indicated that this outcome is also attributable to the high degree of in situ
SO2 capture capability of lime kilns (NCASI, 2009a).
2.5.2.3,	Lime Kiln PM Emissions
While passing through the kiln, the combustion gases pick up a significant amount of PM both from lime mud
dust formation and from alkali vaporization (NCASI, 2004). Particulate matter in the exhaust gas is mainly sodium
salts (Na2S04 and Na2C03), CaC03, and CaO (USEPA, 1996).
2.5.2.4,	Lime Kiln GHG Emissions
As mentioned previously, in the kraft pulping and chemical recovery process, biomass carbon from the wood is
dissolved and either emitted as biomass CO2 from the recovery furnace or captured in Na2CC>3 exiting in the
smelt from the recovery furnace. In the process of converting the Na2C03 into new pulping chemicals, this
biomass carbon (i.e., the carbonate ion) is transferred to CaCC>3 in the causticizing process. In the lime kiln, the
CaCC>3 is converted to CaO (i.e., lime material used in the chemical recovery process) and biomass CO2 originating
from the wood residuals contained in black liquor is released to the atmosphere. Unlike lime kilns used at lime
production facilities, where CO2 emissions are entirely fossil in nature, the CO2 emitted from kraft mill lime kilns
originates from two sources: 1) fossil fuels burned in the kiln, and 2) conversion of CaC03 (or "lime mud")
generated in the recovery process to CaO (lime). The CaC03-derived CO2 emissions almost exclusively originate
from biomass and are accounted for in recovery furnace emission factor calculations (because recovery furnace
emission factors are based on the carbon content of spent pulping liquor, and this biogenic carbon eventually
exits the chemical recovery process from either the recovery furnace or lime kiln) (USEPA, 2010a).
The lime kiln typically produces about 95 percent of the lime needed for the causticizing reaction. Either make-
up lime or limestone is purchased to account for losses (USEPA 2010a). Emissions associated with carbonated
makeup chemicals are typically included in emissions inventories. Therefore, the Universal ISIS-PNP does not
include CO2 emissions associated with makeup chemicals.
2-38
-------
Several pulp mills pipe stack gas from lime kilns or calciners to adjacent precipitated calcium carbonate plants
for use as a raw material. Precipitated calcium carbonate is sometimes used as an inorganic filler or coating
material in paper and paperboard products (USEPA, 2010a). This practice is not accounted for in the Universal
ISIS-PNP because it transfers emissions offsite rather than reducing them through implementation of emission
reduction measures.
In addition to CO2 emissions, lime kilns emit CH4 and N2O from combustion of fossil fuels. According to the
International Council of Forest and Paper Associations GHG Emissions Estimation Protocol (NCASI 2005), the
operating temperatures in rotary lime kilns (the predominant design for pulp mill lime kilns) appear to be too
high to allow significant generation of N2O. Therefore, it is reasonable to assume that N2O emissions from rotary
lime kilns are negligible (USEPA, 2009).
2.5,3, Lime I<
2.53.1. Lime Kiln NOx Reduction
This section discusses the applicable controls for reducing lime kiln NOx emissions, based on information from
NCASI 2009a. According to the report, combustion modifications are the best prospect for altering NOx
emissions from lime kilns. The report indicated that NOx control in newer lime kilns may be achieved mainly by
minimizing the hot end temperatures in gas-fired kilns and by reducing the available oxygen in the combustion
zone in oil-fired kilns. However, these combustion-related modifications may be difficult to achieve in certain
existing kilns due to their inherent design and the implications for product quality. The report stated that there
are combustion conditions that must be sustained to efficiently produce an end product (CaO) of consistently
acceptable quality and that implications for adversely affecting emissions of other pollutants also need to be
considered. As a result, attempts to modify NOx formation by adjusting the kiln operating parameters, flame
shape, air distribution and excess oxygen have not been very successful (NCASI, 2009a).
According to NCASI 2009a, NOx control strategies for each kiln must be evaluated on a case-by-case basis
because mechanisms of formation and control are not well understood (NCASI, 2009a). To illustrate, the report
pointed out that techniques to minimize the hot end temperatures in gas-fired kilns, while potentially helpful in
reducing NOx emissions, must be balanced with the simultaneous need to address emission levels of TRS
compounds and to sustain the necessary calcining capacity, which would otherwise be reduced underthis option.
The report stated that there is also an energy penalty associated with the need for greater heat input per ton of
lime mud processed. The report further stated that, while reducing available oxygen in the kiln combustion zone
may be useful for NOx reduction in oil-fired kilns, altering the air supply also affects combustion efficiency,
resulting in excessively high emissions of TRS compounds and CO. Whatever combustion modifications are made
may be limited by kiln configuration and geometry, as well as by impacts on process performance, stability, and
control (NCASI, 2009a).
A report by STAPPA referred to a BACT analysis performed in 2003 for a proposed lime kiln. The BACT analysis
evaluated the feasibility of low-NOx burners, flue gas recirculation, oxidation/reduction scrubbing, SCR, SNCR
and non-selective catalytic reduction, and concluded that none of these options is technically feasible for lime
kiln NOx emissions control. Rather, the regulatory agency concluded that "good design and operation practices"
constitute BACT, and established a NOx limit of 175 ppm at 10-percent oxygen (STAPPA, 2006).
Table 2-13 presents lime kiln NOx control technology options and the impact of this technology.
2-39
-------
Table 2-13. Lime Kiln NOx Control Technologies
Control Option
Description
Performance
Application
Combustion Air
Control
Combustion zone availability of oxygen is a key
factor in NOx formation, especially in oil-fired
kilns. Primary air feed is driven by flame control
requirements, limiting the opportunity for
staging combustion air. Air supply must be
sufficient to sustain oxidizing conditions
throughout the kiln.
NOx reduction is
site-specific since
mechanisms of
formation and
control not well
understood
Detuning a burner from optimized
combustion incurs an energy
penalty by virtue of requiring
greater heat input per ton of
product. Inadequate air supply
contributes to excessively high
emissions of TRS and CO, as well
as excessive carbon deposits in the
lime.
More applicable for new kilns, not
retrofit of existing kilns.
Burner Design
Low NOx burners are technically infeasible due to
complex factors that result in poor efficiency,
increased energy usage, and decreased calcining
capacity of the lime kiln. Reduced flame
temperature, however, could be conducive to
diminished thermal NOx formation, especially in
gas-fired kilns.
LNB infeasible; NOx
reduction unknown
for reduced flame
temperature
option
Reducing flame temperature in
gas-fired kilns can reduce NOx, but
that reduction comes with a cost
of reduced kiln capacity or an
energy penalty associated with the
need for greater heat input per
ton of lime mud processed.
More applicable for new kilns, not
retrofit of existing kilns.
Fuel Selection
Fuel nitrogen is the principal source of NOx in oil-
fired kilns, unlike gas-fired kilns where thermal
NOx formation is prevalent. There is typically
little difference in reported emissions between
oil and gas.
NOx reduction fuel-
specific

Flue Gas
Recirculation
(FGR)
A possibly promising but untested approach.
NOx reduction
unknown; untested
approach for lime
kilns
Altering kiln temperature profiles
with FGR would possibly adversely
affect calcining efficiency.
SCR
Infeasible due to kraft lime kiln configuration.
High particulate loadings preclude SCR prior to
particulate control and temperature
requirements are not met after particulate
control.
Infeasible
Reheating the flue gas after the
particulate control device and
ahead of the SCR section would
incur a substantial energy penalty.
SNCR
Infeasible due to kraft lime kiln configuration.
The necessary elevated temperature regime
required for SNCR is unavailable in kilns.
Infeasible

Note: all data from (NCASI 2009a).
2,5,3.2, Lime Kiln SO2. Reduction
According to NCASI 2009a, combustion modifications, as a practical matter, provide little opportunity for
beneficial reduction of SO2 emissions originating in fuels or raw material (lime mud) fed to the kiln. However,
the regenerated quicklime in the kiln acts as an excellent in situ scrubbing agent for reducing SO2 emissions.
Post-combustion controls can provide additional SO2 control. Nearly 70 percent of lime kilns in the US are
equipped with wet scrubbers, over 20 percent are equipped with ESPs, and less than 10 percent are equipped
2-40
-------
with ESPs followed by scrubbers. While the wet scrubbers are primarily venturi scrubbers designed for PM
control and not normally conducive to the gas absorption needed for reduction of SO2, wet scrubbers can
augment the SO2 removal process because the scrubbing solution becomes alkaline from the captured lime dust.
Between the in situ scrubbing and venturi scrubber, approximately 95 percent of the SO2 formed within the kiln
is captured prior to release, typically resulting in low SO2 emissions (approximately 50 ppm) (NCASI, 2009a).
While most lime kilns are equipped with wet scrubbers, installing ESPs to control PM from lime kilns has become
more widespread in recent years (USEPA, 1996). However, according to NCASI 2009a, emissions of SO2 are
higher when ESPs are used for PM control instead of scrubbers. Sulfur dioxide emissions are also affected by the
relative magnitude of sulfur input to the kiln and the sodium content of the lime mud. The improved collection
of fine PM with ESPs and improved lime mud washing contribute to potentially greater SO2 emissions reductions.
The report pointed out that these examples illustrate the compromises that must be struck to balance
environmentally sensitive manufacturing process improvements with collateral changes in other measures of
environmental interest, as well as choosing among emissions control options that may favor one pollutant over
another (e.g., PM vs. SO2) (NCASI, 2009a).
Table 2-14 presents lime kiln SO2 control technology options and their impacts.
Table 2-14. Lime Kiln SO2 Control Technologies
Control
Option
Description
Performance
Applicability
Fuel
Selection
Using fuel with lower sulfur content
S02 reduction
fuel-specific
S02 formation not only dependent upon
fuel sulfur content but also lime mud sulfur
content and sulfur-bearing NCGs or
stripper off-gases that may be burned in
the kiln.
Scrubber
The majority of kilns are equipped with wet
scrubbers for particulate control. Alkaline
conditions accompanying lime dust capture
contribute additional control of S02 not
otherwise retained within the kiln.
Combined with in
situ S02 removal,
typically, >95 % of
S02 is captured
Particulate scrubbers are designed and
optimized for particulates. Associated high
velocities are not conducive to gas
absorption; S02 removal would not likely
equal what might be achievable with a
scrubber designed for that purpose.
Some kilns use ESPs followed by wet
scrubbers.
Note: all data from (NCASI 2009a).
2,5,3,3, Lime Kiln PM Reduction
The following paragraphs summarize available information on PM control technologies for kraft lime kilns from
USEPA's background document for the chemical recovery combustion sources NESHAP (USEPA, 1996).
Due to State and Federal regulations for PM emissions, all lime kilns are equipped with add-on PM control
devices. As noted previously, as a baseline, nearly 70 percent of lime kilns are currently equipped with wet
scrubbers, over 20 percent are equipped with ESPs, and less than 10 percent are equipped with ESPs followed
by scrubbers. If a wet scrubber is used, a mechanical collector (e.g., cyclone separator) may be installed
upstream. The mechanical collector is generally used to remove larger particles, which are mainly calcium-
2-41
-------
containing (CaC03, CaO). The dust collected by the mechanical collector is returned directly to the lime kiln. A
wet scrubber or ESP follows for removal of smaller particulates, which are mainly sodium salts (Na2S04 and
Na2CC>3) and have aerodynamic diameters less than or equal to 10 urn (USEPA, 1996; RTI, 2013a; NCASI, 2004).
Venturi scrubbers are the most commonly used type of wet scrubberfor lime kilns. Water is the typical scrubbing
fluid, but caustic and weak wash are also used. The scrubbing fluid is recirculated, and the scrubber blowdown
is recycled to the lime mud washer. Venturi scrubbers are designed to remove PM primarily by impaction
through high-energy contact between the scrubbing liquid and suspended PM in the gas stream. A venturi
scrubbing system typically consists of a venturi scrubbing vessel and cyclonic separator. The venturi scrubbing
vessel has a converging section, a throat section, and diverging section (USEPA, 1996).
The performance of the scrubber in terms of PM collection is strongly affected by the pressure drop across the
scrubber throat, the liquid-to-gas ratio, and the particle size distribution. Particulate matter collection efficiency
generally increases as the throat velocity and turbulence of the gas stream increase, as indicated by an increased
pressure drop across the scrubber. For lime kiln applications, PM collection efficiencies for venturi scrubbers
average 99 percent (USEPA, 1996).
Although venturi scrubbers have traditionally been the most common PM control device used on lime kilns, the
use of ESPs to control PM emissions from lime kilns has steadily increased since about 1980. The expected
lifetime of a lime kiln ESP is typically more than 15 years (USEPA, 1996).
The ESP is generally mounted on top of the lime kiln feed building, and the captured dry PM is rerouted to the
kiln by gravity feed. The trend towards ESPs as PM control devices at newer lime kiln installations and as
replacement control devices for older scrubbers is primarily related to the lower energy costs, reduced
maintenance, and increased reliability of the ESPs in comparison to venturi scrubbers that provide equivalent
control. An added benefit is that an ESP installed on a lime kiln produces a dry product that can be recycled
directly to the kiln. The wastewater produced by the venturi scrubber is typically recycled to the mud washers
before the kiln to recover the lime particulate in the spent scrubbing flu id. Additional energy is needed to remove
the excess water in the lime mud filter and to complete evaporation in the kiln. Properly designed and operated
ESPs used on lime kilns routinely achieve PM removal efficiencies of 99 percent or greater (USEPA, 1996).
However, as noted in the previous section, emissions of SO2 are higher when ESPs are used for PM control
instead of scrubbers, illustrating the compromises that must be struck in choosing among emissions control
options that may favor one pollutant over another (e.g., PM vs. SO2) (NCASI 2009a). Based on a review of
available lime kiln PM data that show that ESP-scrubber combinations do not necessarily outperform ESPs on
PM, adding a scrubber to an existing ESP (or vice versa) may not necessarily improve lime kiln PM control (RTI,
2013a).
Potential control technology options for reducing lime kiln PM emissions involve replacing the existing scrubber
with an ESP, upgrading or replacing the existing ESP, upgrading the existing scrubber, and adding a second PM
control device (i.e., adding ESP before existing scrubber, or adding scrubber after existing ESP). Attributes of
potential lime kiln PM control technology options are summarized in Table 2-15.
2-42
-------
Table 2-15. Lime Kiln PM Control Technologies
Control Option
Description
Performance
Applicability
Replacing
Scrubber with
ESP
The use of ESPs to control PM emissions from
lime kilns has steadily increased since the
1980s, due to lower energy costs, reduced
maintenance, and increased reliability; also,
lime kiln ESPs produce a dry product that can
be recycled directly to the kiln. Particulate
matter removal efficiencies of 99 % or greater
can be achieved with properly designed and
operated ESPs (USEPA, 1996).
Could increase PM
control from 99 %
(scrubber) up to 99.97 %
(best-performing ESP)
Emissions ofSCbare higher
when ESPs are used for PM
control instead of scrubbers
(NCASI, 2009a).
ESP Upgrade or
Replacement
Upgrading existing ESP or replacing older ESP
with new ESP.
Could increase PM
control from 99 % up to
99.97 % (best-
performing ESP)
The cost of retrofitting a lime
kiln with an ESP is heavily
influenced by site-specific
factors including the age and
design of the ESP already in
place (NCASI, 2009a).
Scrubber
Upgrade
The majority of kilns are equipped with wet
scrubbers for particulate control, obtaining PM
reductions of approximately 99 % (NCASI,
2009a).
Could increase PM
control from 99 % up to
99.88% (best-
performing scrubber)
Alkaline conditions
accompanying lime dust
capture contribute additional
control of SO2 not otherwise
retained within the kiln.
ESP + Scrubber
An estimated 10 lime kilns are equipped with
both a wet scrubber and ESP for particulate
control, obtaining PM reductions of
approximately 99 % on average (best-
performing unit is 99.8 %) (RTI, 2013b).
Could increase PM
control from 99 %
(baseline scrubber or
ESP) up to 99.8 % (ESP +
scrubber)
However, a review of available
lime kiln PM data shows that
ESP-scrubber combinations do
not necessarily outperform ESPs
on PM, so adding a scrubber to
an existing ESP (or vice versa)
may not necessarily improve
lime kiln PM control (RTI,
2013a).
Note: PM control performance data based on (RTI, 2013b).
2,5,3.4. Lime Kiln CO2 Reduction
The USEPA is presently unaware of control measures to reduce fossil-related GHG from pulp mill lime kilns other
than process changes or energy efficiency measures. Process changes and energy efficiency measures are
described below.
Lime kiln oxygen enrichment. Oxygen enrichment is an established technology for increasing the efficiency of
combustion and has been adopted in various forms by a number of industries with high-temperature
combustion processes (e.g., glass manufacturing). According to one study, oxygen enrichment of lime kilns can
reduce fuel requirements by approximately 7 to 12 percent. Reportedly, capital investments for oxygen
enrichment are negligible compared to other recausticizing plant upgrades, requiring relatively simple
equipment, including feed piping, an injection lance, and controls. Payback periods have been estimated
between one and three years (USEPA, 2010a).
Lime kiln modification. Several other modifications are possible to reduce energy consumption in lime kilns.
High-efficiency filters can be installed to reduce the water content of the kiln inputs, thereby reducing the
required evaporative energy. Higher efficiency refractory insulation brick can be installed to reduce heat losses
2-43
-------
from the kiln. One estimate indicated that newer high-performance refractory can lead to lime kiln energy
savings of up to 5 percent. Heat can also be recovered from the lime and from kiln exhaust gases to pre-heat
incoming lime and combustion air. According to one estimate, the energy savings achievable from implementing
all of these measures is approximately 0.47 MMBtu per ton of production. Such improvements may also improve
the rate of recovery of lime from green liquor, thereby reducing a mill's requirement for additional purchased
lime (USEPA, 2010a).
Lime kiln electrostatic precipitators. Electrostatic precipitators can replace wet scrubbers on lime kilns and lead
to energy and water savings. Electrostatic precipitators can collect kiln dust as a dry material and return the kiln
dust directly to the kiln feed without unnecessarily loading the lime mud filter. In contrast, wet scrubbers require
effluent recycling via the lime mud filter and are significant consumers of water. One estimate indicated that,
for every one-percent reduction in lime mud feed moisture content (through the addition of dry dust), lime kiln
energy consumption is reduced by approximately 46 MMBtu. Another analysis found that increasing mud
dryness from 70 to 75 percent would reduce fuel consumption by 0.4 MMBtu per ton of lime (USEPA, 2010a).
Install a biofuel gasifier to use low-Btu gas in the lime kiln. Biofuels such as hog fuel generated on site can be
gasified resulting in low-Btu gas that can be substituted for fossil fuel in the lime kiln. The gasification process is
usually carried out in a fluidized bed reactor to promote a high rate of heat transfer. The biofuel is injected into
thefluidized bed, where high turbulence causes rapid combustion and gasification of the char. Low Btu gases
generated in the reactor are withdrawn, cooled, scrubbed if needed to remove moisture or pollutants, and can
then be fired in the lime kiln to displace fossil fuel.
Both direct firing of hog fuel (dried to a minimum of 85 % dryness) and hog fuel gasification and low Btu gas
incineration in the lime kiln have been practiced in the Nordic countries since the late 1970s. The technology is
viable, but it may not be economically attractive unless free hog fuel is in excess at the site (NCASI, 2001). A
biofuel gasifier will reduce CO2 emissions from fossil fuel combustion in the lime kiln, replacing them with
biogenic CO2 emissions.
Fossil fuel switching. Similar to fuel switching in boilers, switching from oil firing to natural gas firing in the lime
kiln, or discontinuing use of supplemental fossil fuels with high carbon content (e.g., petroleum coke) could
potentially reduce lime kiln CO2 emissions. Additional literature could be identified and reviewed to determine
the applicability and limitations and C02-reduction potential from fuel switching.
2-44
-------
2.6. References
Andover (2010). Staudt, J. ISIS Emissions Control for Pulp and Paper Plants. Memorandum submitted from
Andover Technology Partners to W. Yelverton, M. Witosky, and E. Torres, USEPA, and K. Hanks, RTI. March 15,
2010.
NCASI (2012). National Council for Air and Stream Improvement. Pulp and Paper Mill Emissions of SO2, NOx, and
Particulate Matter in 2010. Special Report No. 12-03. June 2012.
NCASI (2009a). National Council for Air and Stream Improvement. Environmental Footprint Comparison Tool:
Trade-Offs and Co-Benefits Accompanying SOx and NOx Control. 2009.
NCASI (2009b). National Council for Air and Stream Improvement. Environmental Footprint Comparison Tool:
Effects of Decreased Greenhouse Gas Emissions. 2009.
NCASI (2005). Calculation Tools for Estimating Greenhouse Gas Emissions from Pulp and Paper Mills. Version 1.1.
Prepared forThe Climate Change Working Group of The International Council of Forest and Paper Associations
(ICFPA). Available at: http://www.ghgprotocol.org/calculation-tools/pulp-and-paper. Last accessed July 8, 2005.
NCASI (2004). National Council for Air and Stream Improvement. Compilation of Criteria Air Pollutant Emissions
Data for Sources at Pulp and Paper Mills Including Boilers. Technical Bulletin No. 884. August 2004.
NCASI (2001) National Council for Air and Stream Improvement, Inc. Technologies for Reducing Carbon Dioxide
Emissions: A Resource Manual for Pulp, Paper, and Wood Products Manufacturers. Special Report No. 01-05.
December 2001.
RTI (2013a). Flanks, K. and Flolloway, T. Technology Review and Regulatory Options for the Kraft Pulp Mill NSPS.
Memorandum submitted from RTI to K. Spence and B. Schrock, USEPA. March 29, 2013.
RTI (2013b). Flolloway, T. and Flanks, K. Emissions Inventory for Kraft Pulp Mills and Costs/Impacts of the Section
111(b) Review of the Kraft Pulp Mills NSPS. Memorandum submitted from RTI to K. Spence and B. Schrock,
USEPA. February 4, 2013.
STAPPA (2006). State and Territorial Air Pollution Program Administrators (STAPPA) and Association of Local Air
Pollution Control Officials. Controlling Fine Particulate Matter under the Clean Air Act, September 2006.
Available at: http://www.epa.gov/glo/SIPToolkit/documents/PM25Menu-Final.pdf. Last accessed October 31,
2014
STAPPA (1999). State and Territorial Air Pollution Program Administrators (STAPPA) and Association of Local Air
Pollution Control Officials. Reducing Greenhouse Gases and Air Pollution: A Menu of Flarmonized Options.
Executive Summary. Available at: http://www.4cleanair.org. Last accessed October 1999.
USEPA (2013). US Environmental Protection Agency. Greenhouse Gas Reporting Program data publication page.
Available at: http://www.epa.gov/ghgreporting/ghgdata/reported/pulp-paper.html. Accessed November 8,
2013.
USEPA (2012). US Environmental Protection Agency. Memorandum from J. Bradfield and K. Spence, to Project
Files. Meeting Notes for the December 13, 2012 Meeting with AF&PA and NCASI. December 18, 2012.
2-45
-------
USEPA (2010a). US Environmental Protection Agency. Available and Emerging Technologies for Reducing
Greenhouse Gas Emissions from the Pulp and Paper Manufacturing Industry. Available at:
http://www.epa.Rov/nsr/RhRdocs/pulpandpaper.pdf. Last accessed on October 31, 2010.
USEPA (2010b). US Environmental Protection Agency. AP-42, Fifth Edition. Volume I: Stationary Point and Area
Sources, Chapter 1: External Combustion Sources, Section 1.3: Fuel Oil Combustion. May 2010 (corrected).
USEPA (2009). US Environmental Protection Agency. Technical Support Document for the Pulp and Paper Sector:
Proposed Rule for Mandatory Reporting of Greenhouse Gases. Available at;
http://www.epa.aov/climatechanae/emissions/archived/aha tsd.html. Last accessed February 11, 2009.
USEPA (2003). US Environmental Protection Agency. AP-42, Fifth Edition. Volume I: Stationary Point and Area
Sources, Chapter 1: External Combustion Sources, Section 1.6: Wood Residue Combustion in Boilers. September
2003 (update).
USEPA (2002). US Environmental Protection Agency. USEPA Air Pollution Control Cost ManualSixth Edition.
Section 6: Particulate Matter Controls, Chapter 3: Electrostatic Precipitators. Publication No. USEPA/452/B-02-
001. January 2002.
USEPA (1999). US Environmental Protection Agency. Nitrogen Oxides (NOx), Why and Flow They Are Controlled.
Publication No. USEPA/456/F-99-006R. November 1999.
USEPA (1998a). US Environmental Protection Agency. AP-42, Fifth Edition. Volume I: Stationary Point and Area
Sources, Chapter 1: External Combustion Sources, Section 1.4: Natural Gas Combustion. July 1998.
USEPA (1998b). US Environmental Protection Agency. AP-42, Fifth Edition. Volume I: Stationary Point and Area
Sources, Chapter 1: External Combustion Sources, Section 1.1: Bituminous and Subbituminous Coal Combustion.
September 1998.
USEPA (1996). US Environmental Protection Agency. Technical Support Document: Chemical Recovery
Combustion Sources at Kraft and Soda Pulp Mills. October 1996.
2-46
-------
3. Univer	sal ISIS-PNP Modeling Fr	amewor	k
3,
Universal ISIS-PNP is a sector-based linear programming model that can help analyze optimal pulp and paper
sector operations for meeting demand and pollution reduction requirements over specified time periods. The
objective in a Universal ISIS-PNP simulation is to maximize total surplus over a time horizon of interest for the
pulp and paper sector. The total surplus (consumer surplus plus producer surplus) concept has long been a
mainstay of social welfare economics because it takes into account the interests of both consumers and of
producers (Nordhaus and Heyden, 1977).
Price
Supply Curve
Consumer
Surplus
Po
Equilibrium
Producer
Surplus
Demand
Curve
Di Do
Demand
Figure 3-1. Total Surplus in a Market
In a market at competitive equilibrium, the total surplus can be thought of as composed of producer surplus and
consumer surplus. As shown conceptually in Figure 3-1, the producer surplus corresponding to a quantity Q of
a commodity is the difference between the gross revenue and the inverse supply curve (blue area). Gross
revenue is the result of the price and the quantity consumed. Similarly, the consumer surplus corresponding to
a quantity Q is given by the area under the inverse demand curve up to that quantity minus the gross revenue
(pink area).
It is evident from Figure 3-1 that the total surplus is maximized exactly when quantity Q is equal to the
equilibrium quantity Qe. When the quantity consumed is less than the optimum Qe, e.g. Qi, the consumer pays
a higher price C, resulting in reduction in consumer surplus. At a new equilibrium, the marginal total cost will
increase from Pi of the base case to Ci at the lower quantity Qi. The total surplus shrinks, as depicted by the green
area in Figure 3-1. The producer surplus changes from the equilibrium case as a portion of the total surplus.
Flowever, the framework of the Universal ISIS-PNP does not proportion the total surplus into consumer and
producer surplus. Instead, Universal ISIS-PNP calculates the total surplus as the difference between the total
area under the inverse demand curve less the area under the inverse supply curve.
3-1
-------
Universal ISIS-PNP utilizes the general concept of "spatial price equilibrium" in a network when analyzing the
balance of supply, demand, and trade. In spatial price equilibrium network models, interregional economies are
simulated by finding the balance of demand, supply, and trade that will result in competitive market equilibrium
among the regions.
3.2. Objective Function
The objective function of Universal ISIS-PNP is to maximize the total surplus for the pulp and paper sector over
the selected time horizon. By the equivalent function, the objective is to minimize the total discounted cost over
the same time horizon while meeting the demand. Components of total cost include production cost,
transportation cost, import cost, control cost and energy efficiency cost, as well as emission charge. Each
element of cost is corrected to net present value (discounted) by applying a discount factor for each year within
the time horizon based on a user supplied discount rate.
Individual cost elements of total cost are treated by Universal ISIS-PNP as described below:
1	Production cost - Obtained for each pulp and paper production unit. The production cost of each unit takes
into account the factor input costs of raw material, labor, energy, and other cost components.
2	Transportation cost - cost of transport from supply center to the demand center. Production from each
supply center may be transported to any demand center. Distance from each supply centerto each demand
center is incorporated into the industry inputs.
3	Import cost - calculated by multiplying the quantity of imported goods by the import price for each country
of origin and adding any handling and other associated costs. All imports arrive at the import terminals and
incur transportation costs to reach each demand center; distances from import terminals to each demand
center are incorporated into the industry inputs.
4	Control and energy efficiency costs - Include the capital and variable costs of installing controls and energy
efficiency options to achieve any emission reduction targets governed by the constraints.
5	Emission charge - Added if any allowance price is given for the pollutants.
Thus, the objective function is defined as follows (Eq. 3-1):
3-2
-------
Minimize z = ^ discount factor(time) • production cost(time, production unit) +
y discount factor(time) • transportation cost(time, demand center) +
t,dc
y discount factor(time) • import cost(time, import district) +
t,id
y discount factor(time) • control cost(time, production unit, control option) +
t,i,k
y discount factor(time) • energy efficiency cost(time, unit, efficiency option) +
where:
z is the additive inverse of the total surplus,
t is the time period of interest,
i is the production unit of interest,
id is the import district of interest,
k is the control option of interest,
ee is the energy efficiency measure of interest, and
/ is the fuel of interest
Total cost approximates the inverse supply curve by filling demand from the lowest cost product through
consecutively higher cost products until demand is satisfied. Demand is satisfied when the demand price no
longer exceeds the supply cost. The user chooses a range of interest centered on the expected demand for
demand center and production year, model default is 0.5 to 1.5 times the expected demand. Demand in this
range is divided into a user defined number of steps or intervals; the model default is 100 steps. The inverse
demand curve is used to determine the demand price at the midpoint of each demand step using a constant
elasticity of demand model for each region (Eq. 3-2):
D is the demand for the product with corresponding price P(D),
a is the elasticity of demand relative to price, and
DO and P0 are the initially-specified demand quantity and price, respectively
The total surplus is calculated based on a constant elasticity of demand model in the stepwise integration fashion,
as illustrated in Figure 3-2. The surplus within the demand range considered by the user, from Dmin to the final
demand quantity, is estimated by the product of price at the midpoint of each step and the width of the step.
The benefit associated with demand from zero to Dmin is estimated by the product of Dmin and the demand price
of the first step of the range.
t, i, ee
y discount factor(time) • allowance price(time, fuel) • total emissions(time, fuel)
if
[discount factor(time) • total benefit(time, demand center)]
(3-1)
(3-2)
where:
3-3
-------
Price
P(D)
wid
wid
Quantity
min
max
Figure 3-2 Stepwise Integration of the Inverse Demand Curve
3.3. Production and Costs in Universal ISIS-PNP
The demand for a product in a market can be satisfied by the sum of domestic production (sum of PRQunit) and
foreign imports (sum of IPRierminais) decreased by the amount of exports (sum of EPRierminais), as shown in
Equation 3-3:
IPRxerminals ^' EPRTerminals — i DEMANDsteps	(3_3)
Unit	Terminals	Terminals	Steps
The domestic production capacity changes (to satisfy the demand) take place in the Universal ISIS-PNP modeling
framework based on the treatment of total production-related costs. Total pulp [ZPULPcos ) or paper
[Zpapercost) production-related costs are a sum of production costs (Zp ), imports costs (Zi ), export costs
(ZEc t), associated export/import transportation costs [ZTcost), and recycling cost of paper products {ZRcCost)>
as shown in Equations 3-4 and 3-5 below:
Zpulp cost ~ ^ Upcast + ^ Cost + ^ ZE Cost + ^ ZT Cost	(3-4)
ZpAPERcosT ~ ^ Cost + ^ ^I Cost + ^ Cost + ^ Cost + ^ Cost
Production-related costs include capital costs as well as fixed and variable costs. Fixed costs are in the form of
capital recovery costs. Capital recovery usually depreciates the cost of a new production capacity over the
economic life of the additional capacity using a user defined interest rate for capital expenses. Flowever, user
has the option to add fixed costs, if any, for the existing production units. Variable production costs include raw
material costs, labor costs, operation and maintenance cost, fuel costs, water costs, wastewater costs, electricity
3-4
-------
consumption costs, solid waste costs, controls costs, and energy efficiency costs provided by the user on a unit
of production basis.
Universal ISIS-PNP includes constraints for ensuring that production capacity changes occur in a realistic way.
Production is modeled for five types of units: existing production units, expansion units, replacement units,
projected units and new production units. Existing production units are units currently installed and capable of
producing product. Expansion units are the units associated with increasing production capacity at an existing
production unit. Replacement units are production units built to retire existing production units and replace with
new production units. New production and projected units represent entirely new production capacity.
Production capacity changes occur in Universal ISIS-PNP by analysis of production-related cost components.
These production-related cost components are explained in more detail below.
Electricity production/consumption-. Heat produced from boilers can also be used for electricity generation which
can be used to satisfy a mill's own electricity demand or be sold to a grid. A modern kraft pulp mill is more than
self-sufficient in its electrical generation and normally can provide energy for use by other industries or to the
local community. Thus, electricity consumption/production and its cost are formulated in Universal ISIS as
follows (Eq. 3-6 and 3-7):
ELunit = £ PRQunit,product * {(ELCintensity)
product
(3-6)
— (ELPintensity)}
Zelc ~ ELCCost * ^' ELunit	(3-7)
unit
where ELunit is the electricity consumption/production (kilowatt hours, kWh) per unit, ELCintensity is the
electricity intensity consumed (kWh/ton of product) by a mill, ELPintensity is the electricity intensity produced
(kWh/ton of product), and PRQunit,product's the production quantity (tons of product) of all products in a pulp
or paper mill. Thus, ELunn is the net demand for electricity in a mill. When a mill produces more electricity than
it needs to meet its demand, ELunit becomes negative and represents a profit for a mill. Total electricity
cost, ZELC ($) is calculated by multiplying cost per unit, ELCCost ($/kWh) by total kWh of electricity used by a
unit, ELunit across the entire sector.
Heat production-. Heat in the form of steam is required for pulp and paper production. Two main sources can
provide heat: power boilers and recovery furnaces. Burning black liquor in recovery furnaces supplements the
heat produced with fossil fuel-fired and/or wood-fired power boilers. The heat requirement is modeled as
follows (Eq. 3-8):
H E ATRe q u ;;r e ci f unj t j ^ El unit,pro duct * PRQunit,product ^ PRQunit,product * ^ unit * ^Factor	(3-8)
product	product
where EIunitjProduct is the energy intensity required to produce one ton of product (MMBtu/ton).
PRQunit,product 's the production quantity of all products in a mill. ELCq is the electricity production
quantity from each pulp or paper mill. CFactor is the conversion coefficient to convert electricity to heat. For a
paper mill, the total heat production is the sum of heat required for production and heat required for electricity
generation from power boilers (Eq. 3-9):
3-5
-------
HEATrf + HEATpwRgLR HEATfteqUired(Pulpunit)	(3 9)
For a pulp mill with a recovery furnace, the heat required for pulp and electricity production is the sum of heat
produced from a recovery furnace burning black liquor and heat produced from power boilers. Approximately
70 percent of products (130 % of black liquor per ton of product) are assumed to be produced from black liquor
and burned. Because of no cost for black liquor, the Universal ISIS-PNP uses 100 percent of the heat from the
recovery furnace and produces the rest of the required heat from the power boilers (Eq. 3-10):
Zfuel =	* FUELcost	(3-10)
Lime mud, referred to CaCC>3 precipitate, is produced in causticizing and calcining processes in a pulp mill. The
production of lime mud and emissions is formulated as follows (Eq. 3-11):
LMQ = ( I PLproduct) * LMintensity	(3-11)
product
where LMIntensity is the lime mud intensity from each product (tons lime mud/ton product) andPLproduct
refers to the quantity of a product (tons) produced by a mill. LMQ is the total production quantity of lime mud
in tons from a pulp mill. Emissions quantities are calculated by multiplying of emissions intensity (lb/ton lime
mud) with mud quantity in tons.
Waste paper products are recycled and used as feedstock to produce paper products. Increasing the use of
recycled paper product has continually reduced the environmental impacts of the sector. The recovery rate of
boxboard and other board (BXT) is 91.2 percent, container board (CNT) is 50 percent, corrugating medium (COR)
is 91.2 percent, coated printing and writing paper (CPW) is 56.8 percent, newsprint is 73 percent, packaging and
industrial paper (PIP) is 50 percent, and uncoated printing and writing paper (UPW) is 56.8 percent. In the
Universal ISIS-PNP, mills (facilities) are classified according to whether or not they could purchase recycled fiber
or purchase both market pulp and recycled fiber. Total recycling pulp is calculated based on purchase of recycled
fiberfor each unit (Eq. 3-12):
PLPrbcyCLE = ^ PR Qunit,product * PLPrep	(3-12)
product
where PRQunit,prodUCt is the production quantity of each product and PLPrep is the percentage of recycled
fiber purchased to manufacture each product. Recycling fiber transportation cost is calculated similarly to
domestic transport. Three recycle fiber collection locations are assumed in the United States and collected fibers
are assumed to be transported to a products supply center (SC).
Domestic Transportation: Paper products are transported from SC to demand center. However, all paper
production mills are grouped in three supply centers, North, West and South. Similarly, on the domestic demand
side, Universal ISIS-PNP considers domestic demands from three demand centers (North, South, and West)
within the United States. Paper transportation costs are the costs associated with moving paper products from
paper mills to the demand centers. For pulp transportation, all pulp production mills are grouped into three pulp
supply centers (PSCs) in the United States. Pulp transportation costs are the costs associated with moving pulp
products from PSCs to SCs. Unit transportation costs for each product is a function of location of the supply and
demand centers. Domestic transportation costs between the supply and demand centers are adopted from the
North American Pulp and Paper model (Ince et al., 1994). Each facility (regardless of its regional location) is
allowed to transport paper products to any of the demand centers. Flowever, the model determines the optimal
3-6
-------
transportation route between supply centers and demand centers depending on the transportation cost. Total
cost of domestic transport for pulp products is expressed as (Eq. 3-13):
^T Cost -1 TQPLpC * PLQpC
(3-13)
PC
where TQPLSC is the cost ($/year) of transporting a unit of domestic quantity from each pulp supply center to
each paper supply center. PLQPC is the pulp quantity (tons/year) transported from each pulp supply center to
each paper mill. For the integrated pulp and paper mill, the ZTcost equals zero. Pulp quantity at a pulp supply
center is the total sum of pulp products produced from each pulp unit (tp) in that region and is expressed as
(Eq. 3-14):
TPLpC = ^^PLQip	(3-14)
ip
Similarly, total cost of domestic transport for paper products is expressed as (Eq. 3-15):
ZTcost> = ^ TPRsc * PRQsc	(3-15)
sc
where TPRSC is the cost ($/year) of transporting a unit of domestic quantity from the paper products supply
center to domestic demand centers. PRQsc's the paper products quantity transported from each paper supply
center. Paper products quantity at supply center is the total sum of paper products produced from each paper
unit (ir) in a region and it is expressed as (Eq. 3-16):
TPRSC = ^™2ir	(3-16)
ir
Imports: Import costs (ZIcost) to each import district are the product (of pulp or paper) of the imported quantity
and the cost of importing product. The quantity imported to each import district is iteratively determined from
the marginal cost of domestic production, at the high cost production facility, and the total cost associated with
the imports inclusive of transportation. Cost of imports is the sum of the import price to the import district,
insurance and freight to the import district and handling costs at each import district. The import price is
determined from a constant elasticity of supply curve for each import district based on user supplied information.
In the current version of the model, the import cost function includes two origins of imports (Canada and the
Rest of the World) and three import districts within United States (North, West, and South). Total cost of import
is expressed as (Eq. 3-17):
Zicost =^/(?id * IC0STid	(3-17)
id
where ICOSTid is the import cost including import cost (from Canada and ROW), insurance, freight, customs,
and handling, and associated cost of transporting a unit of imported quantity from the import districts (id) to
pulp or paper demand centers. IQid is the quantity imported from each origin of import to each import district
(id).
Exports: Pulp and paper product are exported to Canada and the Rest of the World from each demand center
and pulp supply centers. Exports quantity is assumed to be the average exported quantity for the last ten years,
3-7
-------
and it is assumed to be terminal capacity. User can define yearly quantity increase percentage (e.g., 3 % per year,
etc.) Exports cost (ZEcost) to each export district is the product of exported quantity and the cost of exporting
product. Costs of exports is the sum of the export price to the export district, insurance and freight to the export
district and handling costs at each export district.
Controls: A controls database provides information regarding applicable air pollution control technologies and
their cost and emission control characteristics. In general, the costs associated with controls comprise the
following components: (1) capital and fixed operation and maintenance costs, (2) costs associated with any
reagent and/or catalyst consumption, (3) costs associated with any reduction in fuel and/or raw material use,
(4) costs associated with electricity consumption, (5) costs associated with byproduct(s), and (6) costs associated
with water use, if any. Various cost elements are escalated appropriately to use values in years of interest (Eq.
3-18 and 3-19):
^Qinit —	(CCpollutant * ^'^QpoIIutant)	(3-18)
pollutant
-I
'-'Cc / CC-, i
(3-19)
where CCunn is the control cost ($/year) of installing controls on an industrial boiler, CCpouutant ($/ton
pollutant) is the cost of control per ton of the pollutant it is controlling, POLQpouutant is total tons of pollutant
produced by a unit. Note that ZCc is total control costs ($/year) for the entire sector calculated by adding
CCUnit (control costs) of each unit.
Energy Efficiency Costs: Some of the most commonly used measures include good O&M measures, air
preheaters and economizers, boiler insulation, minimization of inleakage, and steam line maintenance. The
majority of measures are common, such as burner retrofit capable of substantial CO2 emission reduction. For
example, the replacement of conventional LNB with ULNB is capable of reducing NOx and CO2 emissions by 75
percent (NCASI, 2009) and 6 percent (USEPA, 2010), respectively, compared to uncontrolled case.
The cost elements associated with energy efficiency measures are specific to each mill and individual costs to
upgrade a mill (Eq. 3-20). Zupgrade unit can be added to estimate total cost, Z (Eq. 3-21).
Eupgrade, unit = E0id, unit 0- ~ Percent reduction)	(3-20)
Z — ^ ' -^upgrade, unit	(3-21)
unit
where Eupgrade is emission of a unit after energy efficient measures have been taken, and E0idi unjt is the emission
of a unit before the measures.
3,4. Modeling Framework Architecture
The Universal ISIS is developed in GAMS language and has a modular architecture as shown in Figure 3-3 forthe
PNP sector.
The inputs from Universal ISIS-PNP are transmitted to the optimization part of the Universal ISIS via the interface,
where the inputs from Universal ISIS-PNP are used to solve the selected cases. The Universal ISIS-PNP interface
is a single personal-computer-based executable tool that provides a user-friendly tool for exploring and
3-8
-------
comparing various scenarios of meeting product demand and pollution reduction requirements for an industrial
sector of interest over specified time periods. The Universal ISIS interface allows the user to develop, edit, or
delete scenarios for an industrial sector of interest. The functionality of the interface ensures that users are
allowed to use individually chosen general inputs as well as policy inputs and are able to access the USEPA-
hosted database and Universal ISIS optimization engine to produce output for the desired type of analysis for
the industrial sector of interest (in this case the PNP sector). The interface is a web-based application and is
programmed in C++ Builder and web development programming software in a graphical web-based layout. The
features of the user interface include pull-down menus, mouse support, and point click activation of many of
the features. The Universal ISIS database is fully secured and protected, so that each user's scenario option will
be evaluated individually.
PULP & PAPER
SECTOR
i
UISIS
OPTIMIZATION
ENGINE
UISIS
OUTPUTS
1111L
Figure 3-3. Modular Architecture of Universal ISIS-PNP
The Universal ISIS-PNP interface communicates with the Microsoft SQL database, generates input data sheets,
and transmits to the Universal ISIS-PNP for optimization. The general input interface helps the user to
develop/modify the required modeling framework of the industrial sector of interest which includes time
horizon (simulation period) to be used forthe model runs, reference year, discount rate, time blocks, commodity
characteristics , emission types, fuel types, plant types and characteristics, as well as imports and exports. The
interface allows users to define policy- related parameters such as number of mitigation options, types of policy
pollutants, emission reduction targets and emission reduction percentages, etc. The user can specify the
emission reduction percentage of interest, allowance, banking or non-banking, taxes, minimum reduction levels
and policy horizon (time period) to be used for the model runs. Selected data are then pre-processed in the
Universal ISIS model to arrive at suitable input parameters for use in equations. After pre-processing the data,
Universal ISIS solves for the appropriate levels of production, imports, and controls required for meeting the
constraints associated with commodity demand and emissions, while maximizing total surplus. The outputs are
transferred into a Microsoft SQL database. The interface then helps the user to interpret these outputs in the
desired format (tables, graphs, etc.).
Architecture of an individual module is shown in Figure 3-4 below, using the example of the Universal ISIS-PNP.
Figure 3-4 shows inputs containing industry-specific data, market data, and optimization parameters. The
industry-specific input data characterize unit-level production, capacity, production cost, capital cost, as well as
3-9
-------
fuel types and cost. Input data also provide information about emissions sources, mitigation technologies,
energy efficiency measures, and emissions/fuel intensities. The market input data consist of historical and
projected nationwide consumption, discount rates, cost of electricity, escalation rates, economic life of
technologies, and import and export quantities and prices. Data related to optimization parameters provide
information concerning emissions caps, emission reduction percentages, taxes, emission abatements, banking
options, and allowance options.
Emissions reduction,
emissions abatements
Emissions banking,
allowance, and taxes
IndusUy (mpuK)
Products, capacity, fuels
i> ij i•! -i-l '•/i IK it'll'
P'tIli h 'ti f i'ii toi
ri\t „ id
t" IS^IC'IV v^tjl . i"
intensities, controls, and
controls costs
Tr.iii	acTi^n,
and transport cost
0

Universal ISIS PNP Optimization
Framework
(Pulp and Paper Sector)

Historical arid projected
consumption,, and price
_y
£
/exports
and price

>¦ ut >t > «| '.rt
ti ipm ir~. t (> iiJ i-r
ptitn l I i mm ill > u . ^ ! _*PHt• n t I n'1n' *
fuels quantity, imports, exports,
allowance price, marginal price,
and average price
ptimized mitigation
options, energy
efficiencies and
emissions
city, (list
Figure 3-4. Input and Output Data Management in Universal ISIS-PNP
3.5. Constraints and Limitations
The Universal ISIS-PNP includes constraints and limitations for ensuring that production capacity changes occur
in a realistic way. Constraints include production, consumption-supply, and emissions. Limitations include
transportation and terminal capacities.
Production of a commodity is limited to the availability of the plants. Plant availability can be restricted by
resource availability such as fuel or raw materials availability and by capacity. For instance, energy consumption
by a plant can only be selected from the fuels available at the location of production. The total supply for each
demand center has to be greater than or equal to consumption in the given time period. Supply can be
comprised of local production, import from other regions, and foreign import. Universal ISIS-PNP provides full
flexibility to determine demand centers, imports and exports terminals, commodities quantity and price, and
associated domestic and exports/imports transportation costs.
Emissions abatement approaches in Universal ISIS are categorized in three abatement approaches: process
modifications and upgrades, raw material and/or fuel substitution, and emission mitigation technologies. For
each emission abatement approach, where possible, information on capital cost, fixed operating cost, variable
operating cost, emission reduction performance for all of the pollutants, impacts on fuel and/or raw material
use, impact on electricity consumption, byproduct generation and cost, and impact on water use is included in
the Universal ISIS. The Universal ISIS-PNP framework includes algorithms to account for tracking multiple
pollutant streams associated with uncontrolled emissions, controlled emissions, pollution prevention from
3-10
-------
process modifications and energy efficiency measures, and any controls-related effects. For a given pollutant,
total emissions have to be limited to emission limits specified by the exogenous policy constraints on emissions.
If the policy being analyzed allows for banking of emissions, then the banking equation enables banking of
allowances for future use.
Transportation of goods and commodities from a supply center is limited by lower of the production capacity of
the supply center and the transportation capacity from a supply center to all demand centers, if specified.
Imports quantity at each terminal is limited by the terminal capacity. However, Universal ISIS-PNP provides full
flexibility to customize assumptions including changes in quantity (e.g., percentage increase per year), changes
in import prices and terminal locations.
The objective function of Universal ISIS-PNP is minimized with regard to the constraints described above to arrive
at the optimal solution.
3,6. Optimization and Post-Processing
In Universal ISIS-PNP, the input data are pre-processed to arrive at suitable input parameters for use in the
model equations explained earlier in this chapter. Once the data have been pre-processed, Universal ISIS-PNP
solves for the appropriate levels of production, imports, and controls required to meet the constraints
associated with product demand and emissions while maximizing total surplus. Once the surplus maximization
problem has been solved, the results are post-processed to obtain parameters and level values of the variables
of interest. The key variables of interest are: production level of each production unit to meet regional demand,
level of imports in each region, installation of various controls, emissions, and various costs. Output data are
written in appropriate worksheets in an Excel workbook and further linked to various plots to enable visual
presentation and analyses of the results. The Universal ISIS-PNP modeling framework is designed to
accommodate the analyses of emission reduction technologies for multiple pollutants. For a particular emission
control strategy under consideration, the Universal ISIS-PNP can estimate the amount of emission reductions
and associated costs. Universal ISIS-PNP may analyze a number of emissions reduction options including fuel
exchange, cap-and-trade, emission taxes, emissions limits, and target emissions reduction. Additionally,
appropriate combinations of these options can also be evaluated.
The fuel switching option offers substantial reductions of emissions from the pulp and paper sector. Fuel
switching is an attractive option for reducing boiler emissions because these emissions are a function of fuel
consumption. For example, combustion of natural gas produces far less SO2 emissions than coal because of its
significantly lower sulfur content. Natural gas and oil are favorable fuels from the standpoint of NOx emissions
compared to coal and wood. These examples of different fuel switching scenarios can be analyzed by Universal
ISIS-PNP and an optimum fuel switching for minimal emissions can be selected.
In the cap-and-trade option, an emissions cap is set on the amount of a pollutant that can be emitted by the
sector considered. Sources are issued emission permits (allowances) that represent the right to emit a specific
amount of the pollutant. Allowances may be banked for use in future. The total amount of allowances available
in the current period and those banked in previous periods cannot exceed the cap in the current period. Sources
or companies that need to increase their emissions may buy allowances from those who pollute less. This
transfer of allowances is referred to as an allowance trade. In effect, the buyer pays a charge for polluting, while
the seller is rewarded for having reduced emissions by more than was needed. Thus, in theory, those that can
reduce emissions least expensively will do so, achieving the pollution reduction at the lowest possible cost to
3-11
-------
the sector. The Universal ISIS-PNP framework allows the user to select an allowance price to determine the level
of emission reduction achieved by the sector corresponding to this selected allowance price. The cost of
emissions is determined for each pollutant as the product of the emission and allowance price for the emission
considered.
Universal ISIS-PNP framework allows also for evaluation of costs and emission reductions associated with
emission reduction programs utilizing unit-specific rate-based emission limits. This is accomplished by imposing
the rate-based emission limit for pollutant emitted by any specific unit.
3,7. References
Ince, P. et al. (1994). Recycling and Long-Range Timber Outlook, Background Research Report, 1993 RPA
Assessment Update, USDA Forest Service. US Department of Agriculture, Forest Service, Forest Products
Laboratory. Research Paper FPL-RP-534. 1994.
NCASI (2009). National Council for Air and Stream Improvement. Environmental Footprint Comparison Tool:
Trade-Offs and Co-Benefits Accompanying SOx and NOx Control. 2009.
Nordhaus, W. and Van der Fleyden, L. (1977). Modeling Technological Change: Use of Mathematical
Programming Models in the Energy Sector, Cowles Foundation Discussion Papers 457, Cowles Foundation for
Research in Economics, Yale University. 1977. Available at: http://cowles.econ.vale.edu/P/cd/d04b/d0457.pdf.
Last accessed on October 31, 2014.
USEPA (2010). US Environmental Protection Agency. Available and Emerging Technologies for Reducing
Greenhouse Gas Emissions from the Pulp and Paper Manufacturing Industry. Available at:
http://www.epa.gov/nsr/ghgdocs/pulpandpaper.pdf. Last accessed on October 31, 2010.
3-12
-------
4. Pulp and Paper	Data
A significant amount of data is required for the Universal ISIS-PNP, including mill-level data, sector data, policy
information, and economic parameters. The Universal ISIS-PNP models the pulping and papermaking processes
separately to represent the various types of facilities found in the sector (i.e., integrated facilities, non-integrated
pulp mills, and non-integrated paper mills) more accurately. The Universal ISIS-PNP focuses on the combustion
sources located at pulp and papermaking facilities. These combustion sources include recovery furnaces, lime
kilns, and boilers. Recovery furnaces are used to recover valuable pulping chemicals for re-use in the process.
Lime kilns are used to convert the recovered pulping chemicals into fresh pulping chemicals. Boilers, often
capable of being fired with multiple fuels, are used to generate steam and electricity for the facility and its
processes.
Boilers, recovery furnaces, and lime kilns do not produce paper products but are supporting equipment in a
larger production process that encompasses a variety of equipment. The onsite production of steam for
electricity generation and process heating is a major supporting operation in paper manufacturing. Production
of steam is accomplished by utilizing renewable energy sources, primarily byproducts of wood preparation and
virgin pulping processes. On average, more than 40 percent of electricity is produced onsite along with
cogeneration of steam. Process energy consumption can vary widely due to facility-specific operations. Boilers
are fired with a diverse range of fuels and exhibit varying boiler efficiencies as shown in Table 4-1 (DOE, 2005).
Table 4-1. Boiler Fuel Efficiency
Fuel Type
Boiler Fuel Efficiency (%)
Oil
83
Gas
82
Coal
81
Bark
64
Black Liquor
65
In 2002, the pulp and paper sector generated 51,208 million kWh, which represented 38 percent of total US
industry onsite generation (USDOE, 2005). Table 4-2 lists statistics on fuel and energy use by the pulp and paper
sector in 2000, based on data compiled by an industry trade organization (AF&PA, 2002). In 2000, the energy
use mix was dominated by the use of self-generated renewable energy (56 %) and purchased natural gas (18 %).
4-1
-------
Table 4-2. US PNP Sector Energy Use in 2000
Fuel Source
Billion Btu
% of Total
PURCHASED
Electricity
155,319.80
7
Steam
33,882.90
1.5
Coal
265,800.00
12
Petroleum Products
102,184.20
4.6
Natural Gas
395,611.00
17.7
Other
24,052.60
1.1
Excess Energy Sold
44,836.00

Total Purchased
932,014.50
43.9
SELF-GENERATED
Hogged Fuel
327,359.00
14.7
Spent Liquor (solids)
894,985.90
40.3
Hydroelectric Power
4,989.70
0.2
Other
19,866.50
0.9
Total Self-Generated
1,247,201.10
56.1
The data inputs to the Universal ISIS-PNP can be broadly categorized into the following main components:
•	Finished product data
•	Mill level data
•	Cost data
•	Emissions and controls data
•	Import modeling data
•	Policy and economic parameters
This chapter discusses the data collection methodology as well as the components of Universal ISIS-PNP input
data.
4.1. Data Collection Methodology
Data to assist in characterizing individual facilities and the collective industry were purchased from Resource
Information Systems, Inc. (RISI) forfacilities with available data. The RISI data included information on pulp and
paper production, cost, facility characterization, product composition, and import-export. The data sets
purchased from RISI and a description of each set are given below:
4-2
-------
•	North American Graphic Paper Capacity Report 2011- this report summarizes the current, planned, and
future capacity for the North American printing and writing paper market
•	North American Graphic Paper Historical Data -this report contains 17 years of historical data (annual basis),
including production, consumption, imports, exports, capacity, prices, and costs for the graphic papers,
recovered paper, and pulpwood markets
•	World Recovered Paper Annual Historical Data - this report contains 17 years of historical data (annual
basis) for supply, demand, and price for recovered paper
•	North American Paper Packaging Capacity Report 2011 - this report summarizes the current, planned, and
future capacity for the North American paper packaging market
•	North American Paper Packaging Annual Historical Data - this report contains 17 years of historical data
(annual basis), including production, consumption, imports, exports, capacity, prices, and costs for
corrugated box, containerboard, boxboard, packaging and industrial papers, and recovered paper and
pulpwood
•	World Market Pulp Capacity Report 2011 - this report contains current, historical, and future capacity for
the world paper grade market pulp industry
•	World Pulp Annual Historical Data - this report contains 17 years of historical data (annual basis) for the
world market pulp industry
•	World Tissue Capacity Report 2011 -this report contains current, historical, and future capacity of the world
tissue market
•	United States Mill Asset Database - this database provides process flow diagrams for US pulp and paper
facilities
The 2011 Lockwood-Post, a directory of the North, Central, and South American pulp and paper mills, was also
purchased from RISI. This directory is a compilation of survey information obtained annually from many of the
mills and companies listed in the directory, supplemented with data from other sources. The data in the
directory included facility locations, process types, product types, and production capacities. Only facilities
operating in 2011 and those that closed in 2010 were included in the directory, i.e., facilities that were idle or
closed prior to 2010 were not included and neither were facilities idled in 2011 but expected to reopen in 2012.
Emissions and controls data were obtained from an information collection request (ICR) survey sent to the pulp
and paper industry by the USEPA in 2011. This survey collected information for use in regulatory reviews, i.e.,
the review of the National Emission Standards for Hazardous Air Pollutants (NESHAP) for pulp and papermaking
sources and the NESHAP for chemical combustion sources at pulp mills, as well as the review of the NSPS for
kraft pulp mills. Only major source facilities (i.e., those that produce more than 25 tons per year of hazardous
air pollutants) in operation during base-year 2009 were required to complete the ICR survey. Data for area
source facilities (typically stand-alone paper mills) were not collected.
Historical mill information was obtained from the "Mills Online" database
(http://www.cpbis.gatech.edu/data/mills-online-new) maintained by the Centerfor Paper Business and Industry
Studies at the Georgia Institute of Technology. This database provides historical mill data for all facilities which
4-3
-------
have operated in the US since 1970. The data included mill location, number of products produced, and whether
or not a facility was an integrated facility. The website also tracks company announcements such as capacity
expansions, closures, and acquisitions.
Supplemental data for facilities were obtained from company websites, news websites, and trade organization
websites, including the mill curtailments and closures spreadsheet maintained by the Pulp & Paper-workers
Resource Council (http://www.pprc.info/html/millclosures.htm). These data were used to determine if facilities
were closed and to determine product composition (e.g., recycled material content).
4.2. Finished Product Data
Finished product data were purchased from RISI to provide 10 years of historical information for the Universal
ISIS-PNP. This section discusses the processing of the purchased data and a summary of the final data used for
the Universal ISIS-PNP.
4.2.1. Data Processing
Universal ISIS-PNP modeling efforts were focused on representing the entire population of US integrated and
non-integrated pulp and paper mills and their products. The pulp and paper sector produces a wide variety of
products, e.g., printing and writing papers, sanitary tissue, industrial-type papers, containerboard, boxboard,
newspaper, etc. Therefore, it becomes essential to aggregate products into product categories to make
modeling more manageable and to develop benchmark products that are capable of describing the industry.
Paper products were aggregated into eight major product categories. Similarly, pulp products there were
aggregated into two major product categories: softwood pulp and hardwood pulp. Table 4-3 lists the major
product categories for paper products and pulp products utilized in the model. At the same time, the assumption
was made that products of similar functionality, i.e., belonging to the same major product category, will share
the same demand variables.
The methodology of similar product aggregation into product category is illustrated in Figure 4-1 for the
boxboard group of products.
-.i Siataild
30& MlkCarton & ^od
2996 All Otteu'Ex ports
Boxboard <+-
r

23$: Bteadisd Board
123S U is Isadied Board
¦Klt'JreaGoS.t'MCcfi Roeydbd

Can & Drum
3*%m OZ-A-'LiXi-S
3S%zoO -4 BaaMS-d
Erfe Gypsu -n vts ~S
4&S«Lij)BatiiQ&rd
ISftTube Ci i S Drum
23%. All Othftij'Ea ports
4-4
-------
Figure 4-1. Methodology of Product Aggregation into Product Categories
Table 4-3. Universal ISIS Product Categories for the US Pulp and Paper Market
Paper
Categories
Subcategories
1. Containerboard (CNT)
Bleached Kraftliner
Unbleached Kraftliner
White-top Kraftliner
Semi-chemical Medium
Recycled Medium
Recycled Liner-board
White-top Recycled Liner
2. Boxboard and Other
Board (BXT)
Bleached Boxboard
Unbleached Boxboard
Folding Cartonboard
Liquid Packaging board
Food Service
Folding Boxboard
Other Unbleached Boxboard
Recycled Boxboard
Other Recycled Board
Coated Cartonboard
Uncoated Cartonboard
Gypsum Wallboard Facings
Tube, Can, Core, and Drum
Multi-ply/Multi-furnish Boxboard
White-lined Chipboard
Liquid Packaging Board
Bleached Kraft Board
3. Packaging and Industrial
Paper (PIP)
Kraft Wrapping Paper
Unbleached Kraft Paper
Bleached Kraft Paper
Unbleached Packaging Paper
Bleached Packaging Paper
Specialty and Industrial Paper
4. Corrugating Medium
(COR)
Semi-chemical Medium
Recycled Medium

5. Newsprint


6. Tissue


7. Coated Printing and
Writing Paper (CPW)
Coated Free-sheet
Coated Bristol
Coated Free-sheet including coated Bristol
Coated Mechanical
Coated Groundwood
8. Uncoated Printing and
Writing Paper (UPW)
Uncoated Freesheet
Uncoated Bristol
Cotton Fiber Papers
Uncoated Freesheet including Uncoated Bristol
and Cotton Fiber Paper
Uncoated Mechanical
Uncoated Groundwood
Pulp
1. Hardwood Pulp (HWP)
Bleached Hardwood Kraft
Unbleached Hardwood Kraft
Mechanical Hardwood
2. Softwood Pulp (SWP)
Bleached Softwood Kraft
Unbleached Softwood Kraft
Mechanical Softwood
Boxboard comes in various forms, may be bleached or unbleached, and may also be recycled. Furthermore,
many boxboard products serve a similar purpose but have slightly different characteristics to meet niche needs.
Boxboard, a thick, paper-based material that is generally thicker than regular paper, is used for products such as
4-5
-------
milk cartons, cereal boxes, shoe boxes, orfrozen food packaging. In general, boxboard is used forthe packaging
of non-durable consumer goods. However, the functionality of these different boxboard products, the driver of
consumer demand, remains similar. Therefore, because Universal ISIS-PNP models demand, these similar
boxboard products were aggregated (by weighted mean approach) into one product category to arrive at the
total demand for the product category as well as weighted prices and costs.
4,2.2,
Final product data were collected for 514 facilities and 37 product subcategories. Many facilities produce more
than one subcategory of each grade of paper, and as a result, a total of 908 sets of 10-year data were available
for the eight major paper products and the two major pulp products. The Uncoated Printing and Writing Paper
category had the most information with 211 data sets at 146 facilities and contained subcategories such as
uncoated mechanical, uncoated free-sheet, uncoated ground-wood, uncoated Bristol, and cotton fiber papers.
The Boxboard and Other Board category was the second largest category with 170 data sets for 135 facilities.
The Boxboard and Other Board category contained subcategories such as uncoated carton-board, recycled
board, folding carton-board, food service, liquid packaging, bleached kraft board, coated carton-board, other
unbleached board, and other recycled board. The remaining categories can be found in Table 4-4, which
summarizes the number of data sets available for analysis and the number of facilities producing each product.
Table 4-4. Summary of Facilities Producing Each Major Product Category
Major Category
Number of Data Sets
Number of Facilities
Containerboard
77
73
Boxboard and Other Board
170
135
Packaging and Industrial Paper
143
34
Corrugating Medium
70
66
Newsprint
24
24
Tissue
86
86
Coated Printing and Writing Paper
68
58
Uncoated Printing and Writing Paper
211
146
Hardwood Pulp
20
20
Softwood Pulp
39
39
TOTAL Number of Records
908
N/A
Production for the major paper products for 2000 and 2010 is shown in Table 4-5. All of the products except for
tissue experienced a decline during this time period. Newsprint experienced the largest decline with production
dropping from 7,463 tons to 3,588 tons, a decrease of 51.9 percent. Overall industry production declined from
105,030 tons to 89,939 tons, a reduction of 14.4 percent.
4-6
-------
Table 4-5. Summary of Production by Major Product Category
Major Category
2000 Production
(1000 tons/year)
2010 Production
(1000 tons/year)
Change (%)
Containerboard
26,377
25,556
-3.1
Boxboard and Other Board
17,400
14,826
-14.8
Packaging and Industrial Paper
6,197
5,637
-9.0
Corrugating Medium
10,953
10,415
-4.9
Newsprint
7,463
3,588
-51.9
Tissue
7,359*
7,628
3.7
Coated Printing and Writing Paper
11,126
8,814
-20.8
Uncoated Printing and Writing Paper
18,155
13,475
-25.8
TOTAL
105,030
89,939
-14.4
This value is for 2005, the first year of available data for tissue in the RISI report.
Mill level data utilized in the Universal ISIS-PNP were purchased from RISI, collected through the ICR, and
obtained from researching the internet. This section discusses the processing of the mill level data and a
summary of the final data used for the model. This section focuses on the currently operating US facilities as
well as those that have closed since 2000.The Universal ISIS-PNP does not project new mills or options for the
addition of new production capacity by region because no new facilities have been built since 1990, and the
current trend is to reopen closed mills (i.e., International Paper in Franklin, VA). Each mill facility modeled was
characterized by its location, pulping process, facility equipment availability, annual product capacities, and
retirement information, when applicable. In addition, each facility was characterized by its average variable cost
components. This information is discussed further in the following sections.
4.3.1. Data Processing
The number of mills in the Universal ISIS-PNP database and their locations were determined using the North
American Graphic Paper Capacity Report, the North American Paper Packaging Capacity Report, and the World
Tissue Capacity Report purchased from RISI. The mill types (integrated or non-integrated) were determined
using the ICR and Center for Paper Business and Industry Studies' "Mills Online" database. For mills that are
currently closed, a retirement date was determined using the Pulp & Paper Resource Council's website and the
capacity report (i.e., facilities with "0" capacities were assigned the first year of "0" capacity as their retirement
year). Facilities in the Universal ISIS-PNP database were identified based on a plant type and a Universal ISIS-PNP
ID number. These values were concatenated to form mill identification numbers and were then used to identify
the facilities in each data table (e.g., market pulp production, paper production, etc.).
As mentioned previously, the pulp and paper industry is regional in nature. In Universal ISIS-PNP, each facility
modeled is located in one of three regional markets (USDOA, 1994) and shown in Figure 4-2.
4-7
-------
~ North
OR
MT
ND
MX
WA
XH
5 MA
NY
SD
ID
lWIj MI
IA
PA
NE
IN
DE
,nPc
UT
CO
MO
KS
CA
VA
KY
AZ
NM
OK AR
JSC
TX
~ West
~ South
Figure 4-2. Mill Capacity Regions in the US (USDA, 1994)
Although the US market for pulp and paper products exhibits regional behavior, the Universal ISIS-PNP allows
for all modeled facilities to supply demand in any region, subject to transport costs as discussed later. It is
important to note that the purchased RISI data divided the market into two regional markets (north and south)
as opposed to the three regional markets shown in Figure 4-2. Flowever, considering the variation in regional
behavior, we decided that three regional markets would be a better representation of the US market for pulp
and paper. Data for the mills that belonged to the northwest and southwest regions in the RISI database were
assigned to the mills located in the west region of the Universal ISIS-PNP database. Table 4-6 shows the states
in each Universal ISIS-PNP and RISI region for comparison.
4-8
-------
Table 4-6. Universal ISIS-PNP and RISI Region Comparison
RISI Region North
RISI Region South
State
Universal ISIS-PNP
Region
State
Universal ISIS-PNP
Region
CT
North
AL
South
DE
North
AR
South
IA
North
FL
South
IL
North
GA
South
IN
North
KY
South
MA
North
LA
South
MD
North
MS
South
ME
North
NC
South
Ml
North
OK
South
MN
North
SC
South
MO
North
TN
South
NH
North
TX
South
NJ
North
VA
South
NY
North
AZ
West
OH
North
CA
West
PA
North
NM
West
VT
North
NV
West
Wl
North
UT
West
WV
North


CO
West


KS
West


MT
West


OR
West


WA
West


ID
West


4,3.2, Pro
The United States Mill Asset Database, a collection of process flow diagrams (PFDs) for 231 mill facilities, was
used to extract process and production information. The data extracted included pulp production, product
composition, facility type, whether or not the facility could purchase market pulp and/or recycled pulp, number
of boilers and recovery furnaces, fuels utilized, steam and electricity generation, and electricity utilization. These
data were extrapolated to the remaining 283 mill facilities that did not have PFDs available. Production process
information for facilities without PFDs (e.g., the ability to process recycled fiber) was determined using public
information on company websites.
4-9
-------
Production process characteristics for the facilities such as capacity, product recipe, electricity usage, pulp mill
type, equipment classification, controls, and lime mud and black liquor production were determined using the
PFDs purchased from RISI (the United States Mill Asset Database), the ICR, and company websites. These
parameters were used in Universal ISIS-PNP to represent production processes at the modeled facilities more
accurately.
The final product capacity for each product for each facility was determined using the maximum production
value for the 10-year period for which data were purchased. A facility was assumed to be operating at 85-percent
capacity in the year with the highest production. This value represents the maximum amount of each product a
facility can make without making a process change or upgrading equipment. Capacity for a facility was reported
as "0" for years which production values were "0." An example is shown in Table 4-7 below.
Table 4-7. Production Capacity Example
Values
(in thousands of
tons/year)
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
MAX
Production
5.5
7.5
6.0
8.5
10.5
7.5
5.5
5.0
2.5
0
0
10.5
Capacity
12.4
12.4
12.4
12.4
12.4
12.4
12.4
12.4
12.4
0
0
12.4
For integrated facilities, pulp production and pulp capacity were calculated using the PFDs for those with
available data. For pulp production, pre-digester and post-digester pulp values were extracted from the PFDs
and used to calculate pulping yield. For example, a pre-digester value of 2,400 bone dry short tons (BDST) per
day of softwood chips and a post-digester value of 1,170 BDST per day of softwood pulp indicate a pulping yield
of 48.75 percent. This pulping yield was then used in conjunction with the values in the product summary table
(such as Table 4-7) on the PFD for the amount of wood used for each product to determine the amount of pulp
needed for each product. For example, if a ton of finished product requires 1.411 BDST of wood, the product
requires 0.361 BDST of pulp (1.411 x 0.4875=0.361). This value was then multiplied by the yearly production
value for the product to determine the amount of pulp produced. These calculations were repeated for all
products produced at a mill and then summed to obtain the final pulp production value. Final pulp values were
averaged based on the major product category produced, and tons of pulp needed per ton of product values
were assigned to similar facilities where PFDs were unavailable. Pulp mill capacity was determined using the
maximum production yearand a utilizationfactorof 85 percent, the same method used to calculate final product
capacity in Table 4-8.
4-10
-------
Table 4-8. Example Product Summary Table for an Integrated Facility
Category
Units
Product Grade
PIP
CNT
SWRW
BDST/FST
1.411
1.286
SW Chips
BDST/FST
0.734
0.669
Starch
BDST/FST
0.015
0.010
Hog fuel
BDST/FST
0.165
0.169
Coal
Short ton/FST
0.155
0.159
Oil
BBL7FST
0.309
0.288
Electricity
kWh/FST
428
423
BDST = bone dry short ton FST = finished short ton
BBL = barrel	SW = softwood
SWRW = softwood/roundwood
PIP = Packaging and Industrial Paper CNT = Containerboard
Product recipes were determined using the product summary table on the PFD for facilities where PFDs were
available. These values represent the amount of softwood pulp, hardwood pulp, additives, and recycled pulp in
the final product. In the case of integrated facilities, the pulping yield, as calculated for the pulp mill production
and capacity values, was also used to determine product composition. This calculated value and the values for
SWRW (softwood, roundwood), FIWRW (hardwood roundwood), SW Chips (softwood chips), and HW Chips
(hardwood chips) were used to determine the amount of virgin pulp in the final product. Values for starch and
filler from the summary table were also accounted for to calculate an accurate composition. For non-integrated
facilities, values for pulps and recycled papers from the product summary table (such as Table 4-9) were used
to calculate product composition. All recycled paper values (e.g., de-inked pulp, pulp substitutes, old corrugated
containers, old newsprint, and mixed paper) were summed up to create one "recycled pulp" value for each
product. Market pulp values (e.g., northern bleached softwood kraft pulp and northern bleached hardwood kraft
pulp) were maintained as separate values to determine softwood and hardwood percentages. Average product
compositions were determined based on final product and mill classification (e.g., non-integrated vs. integrated,
recycle mill vs. virgin mill) and assigned to facilities without PFDs.
4-11
-------
Table 4-9. Example Product Summary Table for a Non-Integrated Facility
Category
Units
Grade: UPW
Northern Bleached Hardwood Kraft Pulp
ADMT/FST
0.335
Northern Bleached Softwood Kraft Pulp
ADMT/FST
0.211
De-inked pulp/pulp waste
ADMT/FST
0.094
Pulp Subs
ADST/FST
0.156
Starch
BDST/FST
0.026
Filler
BDST/FST
0.188
Natural Gas
MCF/FST
13.11
Electricity
kWh/FST
610
ADMT = air-dried metric ton	ADST = air dried short ton
MCF = million cubic feet	FST = finished short ton
BDST = bone dry short ton
UPW = Uncoated Printing and Writing Paper
The paper grade recipes were confirmed to match the classification of the facility. For example, a facility
producing boxboard from 100-percent recycled material was confirmed to be represented by a "1" in the
"facility can purchase recycled pulp" column. We also confirmed, for example, that a facility producing tissue
from 50-percent hardwood pulp and 50-percent softwood pulp could produce the pulp (indicated by a "Y" in
the "integrated facility" column) or could purchase the pulp (indicated by a "1" in the "can purchase market
pulp" column) for use in the final product. In cases where these conditions were not met, product recipes were
adjusted accordingly; that is, if a product composition was assumed for a facility, and that facility could not
produce or purchase virgin pulp but could purchase recycled pulp, the product composition was changed to
100-percent recycled. Some facilities produced products from both recycled pulp and from virgin pulp. In these
cases, all products were compared to the mill classification to confirm that the facility was capable of making all
of the products based on their characterization. We also confirmed that all pulp mills had corresponding pulp
production data. For example, an integrated facility producing hardwood pulp and making containerboard from
hardwood and softwood pulp was confirmed to be represented by "Y" in the "integrated facility" column and a
"1" in the "facility can purchase pulp" column. For the integrated mills, we confirmed that hardwood pulp
production and/or softwood pulp production values were available. If all years were reported as zero for both
pulp types, the pulp mill classification was removed.
Pulp mill type was determined for the 163 pulp mills in the database. This information dictated the type of
equipment utilized at the facility, as well as the type and quantity of emissions. Pulp mills were classified as
mechanical or chemical based on the PFDs, ICR information, and company websites. Chemical pulp mills were
further classified as sulfate (kraft pulping) or other.
Equipment information was determined using the ICR, the boiler Maximum Achievable Control Technology
(MACT) database, and the PFDs. For integrated facilities, boilers were assigned to paper mills, while lime kilns
and recovery furnaces were assigned to pulp mills. The average number of boilers, lime kilns, and recovery
furnaces (3,1, and 2, respectively) was determined for integrated facilities with data and were assigned to those
without data. The average number of boilers (2) for stand-alone paper mills was assigned to facilities without
4-12
-------
PFDs. Lime kilns and recovery furnaces were not assigned to facilities using chemical pulping methods other
than sulfate (kraft), unless ICR data showed that the facility had one or both.
Existing controls for the assigned equipment were determined based on the ICR and the boiler MACT database
for those with data available. Boiler controls were assigned based on fuel type and products produced at similar
facilities with data. Lime kiln controls were assumed to be scrubbers for all kilns without data, and recovery
furnace controls were assumed to be dry-bottom electrostatic precipitators (DBESPs), both based on the
representative control for the majority of emission units with data.
Lime mud production and black liquor production were calculated using information collected in the ICR. Values
in the ICR for lime mud production were on the basis of a ton of CaO produced per day, and a recovery rate of
90 percent was assumed to calculate the amount of lime mud burned per day. This value was then combined
with the previously calculated daily pulp production for 2009, resulting in a ton of lime mud produced per ton
of pulp produced value. An average value of 0.270 ton of lime mud produced per ton of pulp produced was
assigned to mills without ICR data (this value was the average of all available data). Black liquor values in the ICR
were based on million pounds of solids per day. This value was combined with the previously calculated daily
pulp production for 2009, resulting in tons of black liquor solids produced per ton of pulp produced. An average
value of 1.49 tons of black liquor solids produced per ton of pulp produced was assigned to mills without ICR
data (this value was the average of all available data).
Landfill data were provided in the summary table of the PFDs in the form of waste ton per ton of final product.
For facilities without a PFD, a landfill value was assigned based on mills with data with of similar mill type and
final product type.
43,3, Boiler and Fuel Characterization
Electricity data were extracted from the boiler section of the PFDs for the available facilities. In this section, the
number of power boilers was indicated for each plant, as well as the amount of electricity generated, used,
purchased, and sold. These values were combined with the daily production values to determine the electricity
values (i.e., generated, used, purchased, and sold) per ton of finished product. Facilities with electricity values
were averaged based on all of the products produced at each facility, as well as the facility type, and then applied
to similar facilities producing similar products without PFDs.
Boiler data were extracted from the boiler section of the PFDs for available facilities. The number of power
boilers and the fuel used for each facility were extracted. Fuel intensities, or amounts of each fuel used per ton
of production were also extracted from the PFD summary table. Some boilers had input fuels that were not
assigned to product intensities. These fuel intensities were not populated; however, the fuel intensities were
included as fuels available for use by the facility. An example of this scenario is shown in Figure 4-3. Fuel
intensities were averaged per product and assigned to facilities without PFDs. Controls and boiler fuel types
were assigned based on the most frequently used for each product in the database. Boiler fuel data as well as
the number of boilers were also extracted from the boiler MACT database and compared to the PFD results.
These additional data assisted in populating values forfacilities without PFDs.
4-13
-------
(I)
Power
BDST/PST 0,2
Coal	ST-'FST	015
Figure 4-3. Example Boiler Data
Iloe
Coal
Oil
Facility fuel availability was assigned to each facility based on fuel intensities from the PFD summary table and
boiler input fuels (PFD and boiler MACT). These values were assigned to facilities without PFDs based on final
products produced and facility type.
The total number of facilities represented in Universal ISIS-PNP was 514, which consisted of 151 integrated
facilities, 12 stand-alone pulp mills, and 351 stand-alone paper mills. Individually, there were 163 pulp mills and
502 paper mills represented in the model. These facilities were located in 43 states, divided into three regions.
There were 63 facilities located in the west region, 160 facilities in the south, and 291 facilities in the north. Of
the 163 pulp mills, 145 were classified as chemical pulp mills and 18 were classified as mechanical pulp mills.
According to the populated retirement dates, 346 of the facilities were operating in 2012.
Facilities were classified as to whether or not they could purchase recycled fiber and produce or purchase market
pulp. Of the 163 pulp mills, 49 mills were able to produce market pulp. Many facilities (322) were able to
purchase recycled fiber, and 193 were able to purchase market pulp. Recovery rates for recycled fiber were
estimated for each major product and are shown in Table 4-10.
Table 4-10. Recycled Fiber Recovery Rates for Major Product Grades
Grade
Recovery Rate (%)
Boxboard and Other Board
91.2
Containerboard
50.0
Corrugating Medium
91.2
Coated Printing and Writing Paper
56.8
Newsprint
73.0
Packaging and Industrial Paper
50.0
Uncoated Printing and Writing Paper
56.8
The electricity data showed that 513 facilities consume electricity. The remaining facility utilized only steam
according to the RISI data. Many facilities (391) were able to produce electricity, and 19 of those facilities
produced a surplus of electricity and sold the excess to the grid. A total of 490 facilities purchased electricity.
4-14
-------
Facilities consuming electricity used an average of 899 kWh/ton product. Facilities producing electricity made
on the average 597 kWh/ton and facilities selling electricity made on the average 829 kWh/ton.
Fuel availability was determined for all facilities so that fuel switching could be utilized as an emission reduction
strategy. Fuels such as coal and natural gas may be available to a facility even if those fuels are not currently in
use. Table 4-11 shows the number of facilities with the option to use each type of fuel assessed.
Table 4-11. Fuel Availability Summary
Fuel
Number of Facilities
Coal
210
Natural Gas
488
Oil
335
Hog
171
Free hog
142
Sludge
34
Tire-derived fuel
19
Pet-coke
14
Boilers were assigned to paper mills and stand-alone pulp mills in the model to prevent duplicate counting. All
of the facilities in the database except fortwo had at least one boiler and a total of 1,196 were characterized for
the model. Of the 163 pulp mills, 128 were assigned at least one recovery furnace and 124 were assigned at
least one lime kiln. A total of 217 recovery furnaces and 162 lime kilns were characterized for the model.
Existing controls were assigned to the boilers, recovery furnaces, and lime kilns based on ICR data and
assumptions previously discussed. As shown in Table 4-12, DBESPs were the most commonly utilized air pollution
control device for recovery furnaces (155 units). Wet-bottom ESPs were second most common, utilized on 40
recovery furnaces. As shown in Table 4-13, scrubber was the most commonly used air pollution control device
for lime kilns, followed by an ESP at 27 units. Boiler controls were specific to fuel type and are shown in Table 4-
14.
4-15
-------
Table 4-12. Recovery Furnace Controls Summary
Control Device
Number of Emission Units
DBESP
155
Wet-bottom ESP
40
DBESP-WPR
16
DBESP and Wet-bottom ESP
1
DBESP and SCBR
1
Wet-bottom ESP and SCBR
1
SCBR
2
DBESP and DBESP-WPR
1
Total
217
WPR = wet particulate matter removal	SCBR = scrubber
DBESP = dry bottom electrostatic precipitator ESP = electrostatic precipitator
Table 4-13. Lime Kiln Controls Summary
Control Device
Number of Emission Units
SCBR
121
ESP
27
ESP and SCBR
8
CYC and SCBR
4
CYC and ESP
1
MC and SCBR
1
Total
162
CYC = cyclone MC = multicyclone	SCBR = scrubber
ESP = electrostatic precipitator
4-16
-------
Table 4-14. Boiler Controls Summary
Fuel
Control Device
Number of
Emission Units
Coal
Scrubber - Electrostatic precipitator
19
(total: 248)
Electrostatic precipitator
107

Fabric Filter
118

Scrubber
1

Furnace sorbent injection - Electrostatic precipitator
1

Fabric Filter - Dry sorbent injection
2

Low NOx burners
27
Gas
Venturi Scrubber
2
(total: 596)
No Control
588

Fabric Filter
1

Cyclone or Multi- Cyclone
5

Low NOx burners
270
Dry Biomass
Scrubber
16
(total: 51)
Venturi Scrubber - Electrostatic precipitator
1

Fabric Filter
1

Fabric Filter - Dry Sorbent Injection
2

Electrostatic precipitator
22

Dry sorbent injection and Electrostatic precipitator
1

Duct sorbent injection and Electrostatic precipitator
1

Dry Scrubber - Cyclone
1

Cyclone or Multi- Cyclone - electrified filter bed
2

Cyclone or Multi- Cyclone
4

Low NOx burners
13
Heavy Liquid
No control
41
(total: 47)
Scrubber
4

Spray dryer absorber
2

Low NOx burners
21
Light Liquid
Scrubber
1
(total: 84)
No Control
81

Electrostatic precipitator
2

Low NOx burners
6
4-17
-------
Fuel
Control Device
Number of
Emission Units
Wet Biomass
Wet Scrubber - Electrostatic precipitator
4
(total: 170)
Scrubber
63

No control
1

Spray dryer - Fabric filter
1

Fabric Filter
2

Electrostatic precipitator
95

Dry sorbent injection - Electrostatic precipitator
1

Dry sorbent injection - Cyclone or Multi- Cyclone
1

Cyclone or Multi- Cyclone
1

Dry Scrubber - limestone injection - Electrostatic precipitator
1

Low NOx burners
48
Cost data were obtained from RISI or calculated for six primary cost functions in the model. The primary cost
functions were raw material cost, maintenance and repair cost, labor cost, fuel cost, electricity cost, and solid
waste disposal cost. The development of these functions is discussed in this section.
4.4.1.	Raw
Hardwood and softwood logs and chips serve as raw materials for integrated facilities and non-integrated pulp
mills, whereas market pulp serves as the feed for non-integrated paper mills. The Universal ISIS-PNP identifies
paper mills and pulp mills as two separate entities, however, regardless of whether or not the facility is
integrated. As a result, the raw material cost was divided between pulp mills and paper mills. Raw material cost
for pulp mills included the cost of raw wood, pulping chemicals and wastewater treatment, whereas paper mill
raw material cost would represent the cost of papermaking chemicals and purchased fiber (i.e., recycled fiber
and market pulp). Raw material costs ($/short ton of finished product) for each major product category were
obtained from the RISI database for a period of 11 years (2000 to 2010), and the final raw material cost for each
major product was calculated by summing the weighted average (based on the capacity in the respective year)
of each year's raw materials cost.
4.4.2.	Maintenance and Repair
Repair and maintenance are required for periodic upkeep of facilities. Maintenance and repair costs ($/short
ton of finished product) for each major product category were obtained from the RISI database for a period of
11 years (2000 to 2010). RISI reported the maintenance costs for integrated mills only. Fifty percent of the
reported maintenance costs were considered to make up the maintenance cost of the stand-alone facilities to
conform to the design framework of the Universal ISIS-PNP. Final maintenance and repair cost for each major
product was calculated by adding the weighted average (based on the capacity in respective year) of each year's
maintenance and repair cost.
4-18
-------
Labor
Labor costs were obtained by adding operating labor cost and mill-salaried labor costs. Labor costs were
calculated based on the data reported by RISI Inc. for a period of 11 years (2000 to 2010). However, the RISI
database represented the labor costs for integrated mills only. Therefore, 48 percent of the total labor costs was
used for the non-integrated paper mills, and 52 percent was used for the non-integrated pulp mills. Final labor
cost for each major product was calculated by adding the weighted average (based on the capacity in respective
year) of each year's labor cost.
4.4.4. Fuel
Coal, oil, natural gas, black liquor, biomass, hog-fuel sludge, bark, and TDF are primary fuels used in the pulp and
paper industry. Coal, oil, natural gas, biomass, hog fuel, bark, sludge, and TDF are largely consumed in the boiler,
whereas black liquor is used primarily in the recovery furnace. The Universal ISIS-PNP database gives a detailed
description of the types of fuels used in each pulp and paper mill in the United States. The fuel database of
Universal ISIS-PNP was constructed based on the information collected from the RISI and ICR databases. Fuel
cost ($/MMBtu) in each state for each fuel type has been collected from US Energy Information Administration
website (http://www.eia.gov/state/seds/sep fuel/html/pdf/fuel pr ww.pdf)
¦ i , Elect)	ioty
Electricity is consumed primarily by the auxiliary equipment and paper machine(s). Integrated facilities often
produce more electric power than required and sell extra electricity to the power grid. The Universal ISIS-PNP
database reported the amount of electricity sold, produced and/or purchased with respect to each mill in the
United States. Electricity cost (cents/Kwh) in each state for was collected from National Public Radio website
(http://www.npr.org/blogs/monev/2011/10/27/141766341/the-price-of-electricitv-in-your-state)
"	5	. /osal
Facilities must dispose of production process waste materials. Many of the facilities dispose of waste materials
by utilizing a private landfill. The amount of waste generated per ton of finished product was calculated as
discussed previously and a value of $50 per waste ton was assigned as the disposal cost.
4,4.7. Transf:	?
In the Universal ISIS-PNP, a domestic transportation matrix was used to describe the costs for transporting pulp
from pulp mills to paper mills, and from paper mills to demand centers (the US market for paper products).
Figure 4-4 illustrates the Universal ISIS-PNP network of domestic transport of pulp from pulp mills to paper mills
in three regions within the United States.
4-19
-------
:;>;J p/j.'s por
\ r-f f IJ 0 "orf'I" j
pfTlp/j;apf"r
\ PPPP. fPonPi)

5 r J \ P 'orrfij


^C/DCPPnufj-,}
P rfjp/Ps \Jp.i'
Mi] J
PPPPPP
Figure 4-4. Domestic Transport of Pulp from Pulp Mills to Paper Mills
Transportation costs were calculated by adapting the methodology for the North American Pulp and Paper
modeling framework (USDOA, 1994). However, transportation costs reported in the United States Department
of Agriculture report were adjusted by the consumer price index published by the Bureau of Labor Statistics
http://www.bls.gov/cpi/ to obtain real 2010 dollar values.
4.5. Emissions and Controls Data
The design of the Universal ISIS-PNP can accommodate any number of pollutants of interest. In the model, each
boiler, recovery furnace, and lime kiln were characterized by their NOx, SO2, PM, and CO2 emissions.
4.5.1. NOx
In the Universal ISIS-PNP, pollution control technologies are normally related to boiler, recovery furnace, and
lime kiln heat inputs and/or furnace gas flow rate. NOx emissions from the pulp and paper industry result
primarily from boilers and recovery furnaces. NOx emission reduction methods in the Universal ISIS-PNP may be
divided into combustion and post-combustion methods. Different types of applicable NOx reduction
technologies for recovery furnaces and boilers available in Universal ISIS-PNP are described briefly below and
summarized in Table 4-15, giving NOx control technologies for different types of fuel. Because of different
designs and types of fuels used, not all controls may be feasible for any combustion source in the pulp and paper
sector, as discussed below.
Recovery furnaces - Recovery furnaces are not amenable to low NOx burners because the fuel is not admitted
in a manner where low NOx burners can be applied. However, recovery furnaces can use air staging techniques
in the form of OFA. This type of staging technique would typically be called a "Quaternary" air system, since air
is already admitted in three stages in many recovery furnaces (primary, secondary and tertiary air). Recovery
furnaces are also capable of using post combustion NOx control methods. However, because of the presence of
alkali compounds that would likely poison the catalyst, for SCR it would be necessary to install the SCR after
removal of catalyst poisons from the gas stream.
4-20
-------
Table 4-15. Applicability of NOx Reduction Technologies
NOx Technology
Byproduct Liquor
Biomass
Coal
Fuel Oil
Natural Gas
LNB
No
No
Yes (PC, 50%)**
No (grate/stoker)
Yes (50 %)
Yes (50 %)
ULNB
No
No
No
No
Yes (75 %)*
OFA
Yes (25%)
Yes (25%)
Yes (25%)
Yes (25%)
No*
FGR
No
No
No
No
Yes*
SNCR
Yes (50 %)
Yes (50 %)
Yes (PC 25%)
Other (50 %)
Yes (25 %)
No
SCR
No
No
Yes (80 %+)
Yes (80 %+)
Yes (80%+)***
RSCR (or tail end)
Yes (75 %)
Yes (75 %)
Yes (75 %)
Yes (75 %)
Yes (75 %)
Lo Temp SCR
No
No
No
No
Yes (90 %)
LoTOx****
Yes (90 %)
Yes (90 %)
Yes(90 %)
Yes (90 %)
Yes (90%)
*OFA is generally not an option on package boilers. Since most gas-fired boilers are package boilers, OFA is not generally applicable. FGR, however, is
used but typically in combination with ULNBs.
"Note: expected percent reduction shown in parentheses.
***SCR is not likely to be used for retrofit of gas fired boilers due to low NOx levels that are achievable with combustion controls. However, SCR is an
option for new installations.
****LoTOx requires a downstream scrubber.
Hog fuel boilers- Hog fuel boilers are typically grate-fired or possibly fluid or bubbling bed and are not amenable
to traditional low NOx burners. Hog fuel boilers are also capable of using post-combustion NOx control methods;
however, for SCR it would be necessary to install the SCR after removal of compounds that could poison the
catalyst.
Coal fired boilers - If firing PC, these boilers can use low NOx burners and OFA. Grate- or stoker-fired boilers
would not use low NOx burners, and would instead use air staging similar to OFA. However, coal fired boilers are
also capable of using any post-combustion NOx control method. SNCR would provide about 25-percent
reduction on PC boilers, while higher NOx reductions approaching 50 percent might be achieved in grate or
stoker or fluid/bubbling bed boilers. In principle, a high dust SCR can be installed on any coal-fired boiler, and
hundreds of utility coal-fired boilers employ SCR in a high-dust arrangement.
Heavy Oil-Fired boilers - Number 6 fuel oil or other heavy fuel oil is burned in many pulp and paper mill power
boilers. These boilers typically can use LNB and OFA and any of the post-combustion control methods. Like the
coal-fired boilers, many oil-fired power boilers at pulp and paper mills are not likely to have adequate space after
the economizer for a typical SCR. As a result, a tail-end SCR may be necessary at these locations if SCR is applied.
Natural gas boilers- These boilers are usually most effectively controlled with combustion controls. Natural gas
boilers may use low NOx burners, ultra-low NOx burners, and sometimes flue gas recirculation for control of NOx.
OFA is not likely to be used since most of these boilers are package boilers. SNCR is not likely to be effective on
natural gas fired boilers due to the low NOx levels on gas fired boilers equipped with low NOx burners. SCR would
not be likely to be retrofit on this application because combustion controls tend to be effective at reducing NOx
emissions to low levels. SCR would likely be installed on most new facilities.
4-21
-------
452. S02
SO2 emissions from a recovery furnace are the product of sulfur in the smelt. SO2 emissions from power boilers,
especially those firing coal or residual fuel oil, are directly related to the sulfur content of the fuel. State of the
art recovery furnaces and power boilers maintain efficient SO2 emission controls. Wet and dry scrubbing
technologies may be applied to control SO2 emissions. Wet scrubbing systems capture SO2 in an aqueous
reaction within an absorption vessel. The wet scrubbing processes most commonly used in industrial boilers
such as these in the pulp and paper sector are Limestone Wet Scrubbing, Ammonia Wet Scrubbing, or Sodium
Wet Scrubbing.
The most common form of dry scrubber is an SDA. In an SDA, hydrated lime slurry is introduced into an
absorption vessel to react with the SO2 to form calcium sulfate and calcium sulfite. In all SDA systems, a
particulate matter control device follows the SDA vessel to capture the solids formed in the SDA. In most cases,
the particulate control device is a fabric filter because the filter cake improves SO2 removal performance of the
system.
4.5,3, expand Energy Efficiency
The net CO2 emissions from boilers and recovery furnaces are emitted only from combustion processes. Net CO2
emissions include CO2 emissions from fossil-fuel sources and biogenic CO2 that is absorbed during biomass
growth. Burning biomass in hog fuel boilers produces CO2 emissions from biomass, which equals the total C02
emissions from the hog fuel boiler minus CO2 absorption during the growth of the biomass. In general, plants
absorb carbon dioxide during their growth (life cycle). Short rotation woody crop biomass can absorb 1.88 kg/kg
biomass (0.244 Ib/MMBtu) in its life cycle, which is assumed to be 15 years (Department of Conservation, 2009).
Burning biomass produces 31.34 kg CO2 per kg biomass. Thus, net CO2 emission to the atmosphere is 31.34 -
1.88 = 29.46 kg CCh/kg biomass. CO2 emissions from combustion of fuels used in the pulp and paper sector are
shown in Table 4-16, and resulting emission factors are given in Table 4-17.
Table 4-16. CO2 Production from Combustion of Various Fuels (Ib/MMBtu)

Coal
NG
Oil
Byproduct
Liquor
Biomass
Power Boilers
204.7
123.4
169.1


Recovery Furnace



207.2

Hog Fuel Boilers




241
4-22
-------
Table 4-17. CO2 Emission Factors for Combustion Sources at Pulp and Paper Mills (USEPA, 2009)
Fuel
Emission factor1
(kg CCVMMBtu HHV)
Emission factor1
(lb COVMMBtu HHV4)
Recovery furnaces/black liquor gasification units2
North American softwood
94.4
208.1
North American hardwood
93.7
206.6
North American (average)3
94.1
207.3
Boilers
Biomass (wood and wood residuals)2
93.80
206.8
Coal (mixed - industrial sector)4
93.91
207.0
Natural gas
53.02
116.9
Distillate oil
No. 1 distillate oil
73.25
161.5
No. 2 distillate oil
73.96
163.1
Distillate oil (average)
73.61
162.3
Residual oil (No. 6)
75.10
165.6
Propane
61.46
135.5
Coke oven gas
46.85
103.3
Notes:
1.	Emission factor: to obtain emission factor in lb CCh/MMBtu, multiply emission factor in kg CO2/M M Btu HHV by 2.204623 lb/kg.
2.	Combustion of black liquor in recovery furnaces and black liquor gasification units and combustion of biomass in boilers is considered carbon
neutral with regard to CO2 in GHG reporting protocols. Therefore, CO2 emissions are reported as zero in Universal ISIS-PNP.
3.	The average emission factor for recovery furnace was developed; the same emission factor was used for the black liquor gasification unit.
4.	For coal-fired boilers, emission factors for mixed coals used (industrial sector) were used to reflect mixtures included in pulp and paper inventory.
Replacement of an existing boiler or a recovery furnace is one possibility for improvement of efficiency and
reduction in fuel inputs and emissions outputs as well as O&M costsfor pulp and paperfacilities. Afurnace/boiler
system replacement may require replacing more than the combustion unit to maximize benefit. Because
modern boilers are capable of operating under higher pressure conditions for more efficient steam cycles than
older systems, it is sometimes necessary to replace the steam plant and turbine generator as well as the boiler
to realize the full benefit of the new boiler. Air pollution control equipment will also need to be installed on the
new boiler. Furthermore, a significant improvement in efficiency may result in the paper mill becoming a net
generator, which would require a modification to switchgear and the electrical connection to the electric grid.
Modern recovery furnaces are more efficient than the general majority of existing installed recovery furnaces
due, in part, to the ability of modern furnaces to fire liquor with a higher concentration of black liquor solids
(BLS) and to operate at higher steam pressures and, therefore, more efficient steam cycles. Modern furnaces
also use air preheaters, hot condensate return, flue gas cooling, and other system enhancements that improve
efficiency. Therefore, to realize the full benefits of replacing existing recovery furnaces with new furnaces, it is
generally necessary to modernize the steam plant and to include a new steam turbine generator. The
improvements enable fossil fuel use in the recovery furnace to be relied upon only during startup and shutdown,
thereby increasing power production and reducing CO2 emissions from fossil fuels to near zero.
4-23
-------
Energy efficiency improvement options generally entail an up-front capital cost to install or modify equipment.
The up-front investment is recovered over time through the reductions realized in fuel or electricity costs. In
effect, the variable operating costs are usually negative for these technologies. In most cases, installation of
these technologies will not change the fixed operating costs of the plant. Assuming similar year-to-year facility
operation, the energy savings should be similar from year-to-year. Hence, these methods lend themselves to
being represented economically in terms of simple payback, or how many years of savings required to recover
the cost of the initial investment. A shorter payback period (typically represented in years) implies an
economically more attractive energy savings approach - at least from the perspective of the measure of payback
period.
Capital costs of efficiency improvement methods applied in the pulp and paper industry in 1994 were reported
in terms of $ perton of paper produced per year (Worrell, 2001). Summary of these energy efficiency measures
and their capital costs escalated to 2008 dollars per MMBtu/hour are shown in Table 4-18 along with fuel savings.
Table 4-18. Energy Efficiency Measures for Pulp and Paper Industry Boilers
Energy Efficiency Measure
Capital Cost
($/MMBtu/h)
Fuel Saving (%)
Applicable Share of Production
(%)
Efficient Steam Production and Distribution
Boiler Maintenance
0
6.50
20
Improved Process Control
242
2.8
50
Flue Gas Heat Recovery
424
1.3
50
Blowdown Steam Recovery
484
1.2
41
Steam Trap Maintenance
727
9.2
50
Automatic Steam Trap Monitoring
727
4.6
50
Leak Repair
182
2.8
12
Condensate RReturn
2301
13.8
2
Notes: Capital cost in 2008 dollars. No electricity savings projected for any measure in Table 4-18.
4.6. Import Modeling Data
US pulp and paper markets receive imported quantities of pulp from a number of countries. However, the US
imports a significant amount of pulp and paper products from Canada. Considering Canada's role in the US
import market, the entire import region Universal ISIS-PNP has been divided into two markets: 1) Canada, and
2) ROW. These imports arrive at three import districts: North, South and West. Figure 4-5 illustrates the import
dynamics of Universal ISIS-PNP framework, outlining the import of pulp from world market to the US via import
districts.
4-24
-------
¦jr-r r;
Cf:\" C.f'e*.:.-:',
Figure 4-5. Import Network of Pulp from Canada and/or ROW to the US
4.7, Scenario Parameters
The Universal ISIS-PNP framework allows the user to select a variety of potential emission reduction scenario
options for evaluation. The user can select from cap-and-trade (with or without de minimis requirements),
emissions charge, or rate-based scenario. In a cap-and-trade scenario, separate caps on pollutants of interest
can be specified. The user has the option to run a cap-and-trade scenario with or without banking of emissions.
Further, a cap-and-trade scenario can include de minimis requirements, where the user defines a minimum level
of emission reduction required for each emission unit. As previously mentioned, the user can input an emission
charge for the pollutants of interest. Furthermore, rate-based scenarios with unit specific emission reduction
requirements specified by the user can be modeled in Universal ISIS-PNP. The user can specify the scenario
horizon (time period) to be used for the model runs.
4-25
-------
4.8, References
AF&PA (2002). American Forest & Paper Association, Paper, Paperboard & Wood Pulp: 2002 Statistics, Data
through 2001. 2002.
USDOA (1994). Recycling and Long-Range Timber Outlook, Background Research Report, 1993 RPA Assessment
Update, USDA Forest Service. US Department of Agriculture, Forest Service, Forest Products Laboratory.
Research Paper FPL-RP-534. 1994.
USDOC (2009). US Midwest Average Rainfall, 1971-2000. US Department of Conservation.
www.ncdc.noaa.gov/oa/climate/online/ccd/nrmpcp.txt (Accessed Dec. 5, 2010).
USDOE (2005). Energy and Environmental Profile of the US Pulp and Paper Industry. US DOE Office of Energy
Efficiency and Renewable Energy (EERE), Industrial Technologies Program. 2005.
USEPA (2009) Environmental Protection Agency (2009). Mandatory Reporting of Greenhouse Gases; Final Rule.
Vol. 74, No. 209. October 30, 2009. Office of Air Quality Planning and Standards, Sector Policies and Programs
Division, Research Triangle Park, NC. 2009.
Worrell (2001). Worrell, E., Martin, N., Anglani, N., Einstein, D., Khrushch, M., Price, L., Opportunities to Improve
Energy Efficiency in the US Pulp and Paper Industry, LBNL-48353. 2001. www.researchgate.net Last accessed
November 11, 2014.
4-26
-------
5. Model Calibr	ation
Large techno-economic models of Universal ISIS-PNP framework size require model calibration as they utilize an
extensive amount of data which comes from different sources. This chapter outlines calibration methodology
that was used, discusses data used for calibration, presents calibration results, and gives further
recommendations.
5.1.
The model calibration method utilizes the concept of the calibration constant. The calibration constant has been
developed to account for possible errors in production, imports and costs. The value of calibration constant,
calconst(i), is set by trial and error during calibration. The objective of the trial and error approach is to minimize
the absolute difference in the reported and model-predicted pulp and paper prices (which are marginal values
of the supply equation) for each USGS district.
In the first step of calibration, the Universal ISIS-PNP is set to run for 2007-2009 by making appropriate changes
in the input worksheet and GAMS input files. The import quantities and prices are then adjusted to be equal to
the reported import quantity for each of the import products.
In the next step, the impact of changing the calibration constant is monitored. This impact of the calibration
constant is assessed on estimated production quantities of CNT, BXT, UPW, PIP, and COR products. The
difference between reported and model- predicted production values of all five products should be within
reasonable limits.1 The calibration constant modifies the variable cost of production of each product. The
"Calibration_PnP" worksheet within the "Inputs" workbook has values of the calibration constant assigned for
each product. Finally, in the input GAMS file, the values are assigned for all five products produced from different
units.
The model is first calibrated for the year 2007, to obtain values of the calibration parameter calconst_PnP(i) for
the year. Next, these values are used to validate production and predicted prices against known values of these
parameters for years 2008 and 2009. If there is an acceptable level of difference in the reported and model-
predicted values, the calconst_PnP(i) values for 2007 are used for all future-year predictions. However, if there
is significant disagreement in reported and predicted values, then values for 2007 are used as starting point to
obtain values for the same parameter for 2008. Similarly, the process is repeated to obtain the values for 2009.
Then, an average of the calconst_PnP(i) values over the three years is taken and used for the future model runs.
Current values of the parameter Calconst_PnP(i) being used in the model runs can be found in the worksheet
"Calibration_PnP" of the "ISIS_lnputs.xls" workbook. In the case of the Pulp and Paper model, calibration
constant values for 2007 resulted in an acceptable level of agreement between reported and predicted values
for 2008 and 2009 (discussed in section 5.3, below).
1 There is no standard method to guide the user in determining the acceptable level of "error" in the reported and predicted values
for the purpose of calibration. In this work, we have set an acceptable level for the absolute gap between the individual reported
and the predicted values to ± 15 %, although an effort has been made to keep this level below 10 % for most of the quantities.
However, due to discontinuities in the transportation matrix, errors in the reported data, or other unknowns, the gap in the
estimated and reported values may be higher in certain markets.
5-1
-------
Calibration is a dynamic process; it is recommended that model calibration be performed periodically. In this
fashion, any available new production, imports, or price data could be utilized in the model.
5.2.
Annual production quantities, annual demand, annual imports and reported annual paper prices for the United
States for linerboard, coated free sheet, uncoated free sheet, corrugating medium and solid bleached board
products are the key quantities used for calibration of the Universal ISIS-PNP. Reported data for years 2007,
2008, and 2009 were used to calibrate the model and to obtain values of appropriate calibration parameters.
5.2,1, Prices
Reported annual paper product prices for the United States are shown in Table 5-1. For any given year, the
reported prices for paper products show a wide price-range among product categories.
Table 5-1. Reported Annual Prices of Paper Products
Product
1
£ *c
8 2
™ 55
2008
($/ton)
2009
($/ton)
CNT
495
528
497
BXT
711
745
762
UPW
796
840
822
PIP
659
697
663
COR	464	496	469
5,2.2. Pro
Reported annual paper products production levels are shown in Table 5-2. Generally production has decreased
from 2007 to 2009, due to a decrease in demand resulting from the economic downturn.
Table 5-2. Reported Paper Products Annual Production
Product
2007
2008
2009

Thousand Tons

CNT
20,812
20,329
20,772
BXT
12,910
12,745
12,744
UPW
13,033
12,300
12,244
PIP
2,244
2,113
1,975
COR
9,709
9,561
9,619
5.2,3.
Reported annual linerboard, coated free sheet, uncoated free sheet, corrugating medium and solid bleached
board domestic demand is shown in Table 5-3.
5-2
-------
Table 5-3. Reported Annual Paper Products Demand
Product
2007
2008
2009

Thousand Tons

CNT
20,497
19,973
20,376
BXJ
12,989
12,796
12,768
UPW
17,409
16,675
16,618
PIP
2,878
2,737
2,590
COR
10,047
9,896
9,951
5.2,4, Imp
Reported annual import quantities of paper products to the US are shown in Table 5-4.
Table 5-4. Reported Annual Import Quantities of Paper Products
Product
2007
2008
2009

Thousand Tons

CNT
1089
1089
1089
BXJ
1027
1027
1027
UPW
4424
4424
4424
PIP
941
941
941
COR
443
443
443
5.3. Results
Reported and calculated values of annual prices as well as the difference between the two values forfive paper
products are shown for years 2007, 2008, and 2009 in Table 5-5, 5-6, and 5-7, respectively. As can be seen from
these tables, the majority of differences between reported and calculated product prices are within ±5 percent.
In general, the price differentials are smaller in the year 2007 and have increased in 2009. Generally, model-
predicted market prices are within the criteria specified in the OA document, and the aberrations are explained
by the demand-supply gap and transportation cost.
The current set of calibration constant values are averaged over the years of calibration and are available in the
"Calibration_PnP" worksheet in the "ISIS_lnputs.xls" workbook.
Table 5-5. Reported and Calculated Prices of Products for 2007
Product
Reported
Calculated
%A
CNT
495
488
0.01
BXT
711
705
0.01
UPW
796
757
0.05
PIP
659
708
-0.07
COR	464	451	0.03
5-3
-------
Table 5-6. Reported and Calculated Prices of Products for 2008
Product
Reported
Calculated
%A
CNT
528
500
0.05
BXJ
745
723
0.03
UPW
840
767
0.09
PIP
697
726
-0.04
COR	496	462	0.07
Table 5-7. Reported and Calculated Prices of Products for 2009
Product
Reported
Calculated
%A
CNT
497
489
0.02
BXJ
762
715
0.06
UPW
822
748
0.09
PIP
663
715
-0.08
COR	469	457	0.03
5.4. Recommendations
The Universal ISIS-PNP should be re-calibrated each time modifications or refinements are made. For example,
transportation matrix, modes, and cost of transportation all have significant impact on the behavior of
production distribution and prices across the demand centers. Therefore, following any changes to the
transportation matrix, the model needs to be re-calibrated. In addition, if any of the key input parameters such
as those relating to production quantities and costs are refined or otherwise modified or additional observed
data become available, the calibration of the model should be repeated.
At the time of calibration of the model, production, import, and price values foronly 2007 to 2009 were available.
As discussed above, when the new values become available, the model should be calibrated again. Calibration
of the model should be repeated as soon as new information or new observed data become available. Due to
practical limitations, it is recommended that the calibration of the model be repeated every two years. Further,
since the calibration data are available only for three years, equal weight was given to the parameters obtained
for each year. Once larger data set is available, a modified weighing system can be adopted to give highest
weight to the data from the most recent year.
5-4
-------
6. Illustr	ative Analysis
The Universal ISIS-PNP is designed to show the impact of emission reduction policies on fuel consumption, fuel
and production costs across the pulp and paper sector. The model assists in the analyses of emission reduction
strategies for multiple pollutants: NOx, SO2 and CO2. In this chapter, results of three illustrative scenarios for
reduction of NOx emissions from the pulp and paper sector are presented and discussed. The model-generated
emissions in 2009 are used as reference emissions for each scenario. The emission reduction scenarios
presented here do not reflect any actual USEPA considerations and are used only to familiarize the reader with
the capability of the Universal ISIS-PNP. The following three illustrative examples will be presented below.
1.	Fuel substitution: The first illustrative example for reducing NOx emissions is examined for the year 2010. In
the initial step, the model calculates emissions for the year 2010 based on the emissions in reference year
2009. The emissions cap for NOx is set at 50 percent of the emissions generated by the model in 2010 based
on the reference 2009 year emissions, thereby requiring industry-wide reductions of 50 percent in NOx from
the reference year-based emissions in 2010. The impact of this reduction in NOx is observed on other
pollutants (SOx and PM), fuel switching and fuel costs across the sector. The model chooses from various
fuels of varying cost (coal, natural gas, hog fuel, oil, and black liquor) to minimize operating costs of the pulp
and paper mills while meeting the regional demands and capacity constraints. Low NOx generating fuels
may replace the fuels that emit higher NOx levels. For example, natural gas that has an emission intensity
of 0.19 Ib/MMBtu will replace coal with an intensity of 0.64 Ib/MMBtu.
2.	Installation of controls: The second illustrative example demonstrates how Universal ISIS-PNP can be used
to analyze the effect of installation of an SCR on the operation of a specific pulp and paper plant. The boilers
in some plants are built for specific fuels, thereby constraining them from switching fuels to aid in NOx
reduction. In other words, it may not be possible for some mills to replace their coal fired power boilers
with natural gas fired boilers. In this case, control technologies can be installed on boiler equipment to
reduce NOx emission levels.
3.	Implementation of energy efficiency measures: The third illustrative example describes how Universal ISIS-
PNP could assess the impact of good O&M practices, specifically of replacing or retrofitting burners on
emissions of NOx and CO2. For this example, the same hypothetical plant as the one used in the second
illustrative scenario was used.
6.1. Fuel Substitution
Fuel substitution may be a viable option to reduce NOx emissions from pulp and paper operations because
different fuels have different emission intensities, as shown in Table 6-1, giving emission intensities of different
fuels in Ib/MMBtu.
Table 6-1. Emission Intensity of Fuels

Coal
NG
Oil
Hog
BLS
NOx
0.64
0.19
0.24
0.22
1.40
S02
1.7653
0.00059
0.998
0.025
1.13
C02
208
117
164
4.35
0.75
PM
0.0298785
0.0019
0.062
0.060
0.68
BLS= Black Liquor Solids
6-1
-------
Table 6-1 shows that NOx emissions from BLS and coal combustion are higher than the emissions from natural
gas, oil, and hog boilers. Thus, the use of natural gas offers the opportunity of lowest NOx emissions on an
uncontrolled basis.
Figure 6-1 shows the NOx emissions from the pulp and paper industry in the base case (BAU scenario without
emission reduction requirements) compared with the NOx emissions after applying the 50-percent policy
reduction. The emission reduction requirements may be satisfied either by switching from coal to natural gas or
from coal to oil. The effect of fuel substitution on reducing NOx emissions is also seen on SO2, CO2, and PM
emissions. Figure 6-1 shows that under the emission reduction scenario, the SO2, CO2, and PM emissions are
lower than the BAU case as a result of a different fuel consumption profile. To achieve NOx emission reductions,
the model is constrained to choose a combination of fuels different from the base case. The new fuel
combination ensures that the industry production meets demands with cleaner fuels. Flowever, these new fuels
with lower emission intensities may be more expensive, which may result in an increase in fuel costs and in turn
production costs.
^1980000
39910000
£ 376*0000
130 303
15030J
o 120D03
50303
50303
yUJUJ

| Base Emissions

¦_		
| Coal-Natural gas switcn
¦ Coal-Oil switch

¦ ¦


	^ 		I
COS
NOx	PM
Pollutants
SOh
Figure 6-1. Comparison of Base Emissions with Projected Emissions after Fuel Substitution
The impact of fuel substitution on the annual fuel cost can be seen in Figure 6-2, presenting the modeling results
for base case and 50-percent coal substitution scenarios. The 50-percent substitution of coal with either natural
gas or oil has resulted in a predicted increase of annual fuel cost. Substitution of coal with hog fuel resulted in a
predicted decrease of annual fuel cost.
6-2
-------
3.8E+09
3.7E+09
5 3.6E+09
§ 3.5E+09
o
™ 3.4E+09
(0
> 3.3E+09
Z 3.2E+09
(/)
8 3.1E+09
a)
= 3E+09
2.9E+09
2.8E+09
Basecase	50% of Coal to 50% of Coal to oil 50% of Coal to hog
Natural gas switch	switch	switch
Scenario
Figure 6-2. Fuel Cost in Base Case and Fuel Substitution Scenarios
6.2. Installation of Controls
This scenario illustrates how Universal ISIS-PNP could be used to analyze the effect of installation of an SCR on
the operation of a specific pulp and paper plant. Boilers deployed in the pulp and paper sector typically use LNB
and OFA. However, the boilers are also capable of using any post-combustion NOx control method. In principle,
a high dust SCR (SCR installed upstream of PM control device) can be installed on any industrial boiler. For this
scenario, a hypothetical plant located in the southern supply center with a different annual production capacity
for each of its four products (CNT, newsprint, PIP, and UPW) was selected. The plant uses 6 percent coal, 68
percent natural gas, 11 percent oil, and 15 percent hog fuels to produce the required products. To calculate NOx,
SO2, PM, and CO2 emissions from each of these fuels, the energy intensity (ton of fuel used per ton of product)
and production capacity (tons of product per year) were taken to calculate the amount of fuel (tons of fuel used
per year) used to produce the specific product. The heat produced by each fuel can be calculated by multiplying
the amount of fuel in tons by its corresponding fuel intensity (MMBtu/ton). The emissions (Ib/MMBtu)
associated with each fuel are different based on the emission intensity of that fuel. Each fuel will therefore have
its own NOx, SO2, PM, and CO2 emissions (lb/year). The NOx emissions from all fuel types are added to obtain
total NOx emissions.
To reduce NOx emissions from this hypothetical plant, one could install SCR at the plant. The SCR technology is
capable of 75 percent reduction in NOx emissions from hog fuel and 90 percent emission reduction each from
coal, natural gas, and residual oil. The total NOx emissions (lb/year) as calculated by Universal ISIS-PNP after
applying these percentages of NOx reductions to each fuel results in 85.6 percent overall NOx emission reduction.
However, there is no change in SO2, PM, and CO2 emissions, as can be seen in Figure 6-3.
6-3
-------
675500
¦	Baseline Emissions (no SCR)
¦	Controlled Emissions (SCR)
4500
Pollutants
Figure 6-3. The effect of SCR technology on reduction in emissions of NOx, SO2, and CO2
6.3. Implementation of Energy Efficiency Measures
Energy efficiency measures for industrial boilers used in the pulp and paper sector may vary from operation and
maintenance improvements to repowering. The third illustrative scenario describes how Universal ISIS-PNP
could assess the impact of good O&M practices, specifically of replacing or retrofitting burners, on emissions of
NOx and CO2. For this scenario, the same hypothetical plant, used before in the second scenario (located in the
southern supply center) was selected. The plant uses coal and has conventional LNB installed. This scenario
describes the effect of replacement of conventional LNB with ULNB. The ULNB is capable of reducing NOx and
CO2 emissions by 75 (NCASI, 2009) and 6 percent (USEPA, 2010), respectively, compared to the uncontrolled
case. With conventional LNB installed, emissions of NOx and C02 were assumed to be reduced by 50 and 2
percent, respectively compared to the identical plant without LNB. Baseline emissions of NOx and CO2 for the
plant with LNB were taken as 2,340 and 675,369 metric tons/year, respectively. Next, installation of ULNB was
considered in the third scenario as a replacement for conventional LNB already in place in this plant. As a result
of this LNB-to-ULNB upgrade, emissions of NOx and CO2 decreased to 1,170 and 647,803 metric tons/year for
this plant, as shown in Figure 6-4.
6-4
-------
675000 1
£ 648000 -[
I
c
r 4000
^ 3000
J5 2000
I
1000
0 -
Figure 6-4. The emission reduction of NOx and COa as an effect of LNB to ULNB upgrade.
6.4. Summary
The Universal ISIS-PNP was used to describe three illustrative scenarios of emissions from boilers in the US pulp
and papersector underthe regime of fuel switching, installation of air pollution equipment, and implementation
of energy efficiency measures. The objective of the analysis was to gain insights relative to broad questions on
the range of practical SO2, NOx, and CO2 reduction options in the US pulp and paper industry.
As illustrated by the first scenario, fuel switching offers substantial reductions of SO2, NOx, CO2, and PM
emissions for the sector. Fuel switching is an attractive option for reducing boiler SO2 emissions because these
emissions are a function of fuel sulfur content. For example, combustion of natural gas produces far less SOi
emission than coal because of the significantly lower sulfur content of natural gas. Natural gas and oil are
favorable fuels from the standpoint of NOx emissions compared to coal and wood. As the availability of natural
gas in the US is increasing, more business owners may decide to switch their boilers to natural gas. Universal
ISIS-PNP is a useful tool to predictthe extent of emission reduction resulting from the coal-to-natural gas switch.
As with NOx and SO2, fuel switching is an attractive option for reducing boiler PM emissions. For example, PM
will generally be reduced when a lighter grade of fuel oil is burned or when coal is replaced with natural gas.
Similarly, fuel switching may reduce CO2 emissions significantly because of varying emission intensities of fuels.
For example, CO2 intensity of coal and natural gas is approximately 93 and 53 kg C02/MMBtu, respectively. Thus,
switching from coal to natural gas would accomplish reductions in SO2, PM, and CO2 emissions. Switching from
natural gas to solid biomass would significantly reduce GFIG emissions but would likely increase PM emissions.
These different fuel switching scenarios can be analyzed by Universal ISIS-PNP, and an optimum fuel switching
strategy for minimal emissions can be selected either for an individual boiler or for the sector.
Installation of air pollution control equipment assures reduction of emissions from the plant, as shown in the
second scenario. While over 80 percent emission reduction of a single pollutant from the plant was achieved,
emissions of other pollutants were unaffected. However, Universal ISIS-PNP provides the capability to analyze
emission reduction on a sectoral scale. In this way, the user could implement SCR for plants with the highest
NOx emissions and wet FGD for plants burning fuels with high sulfur content (e.g., high sulfur coal). For example,
utilizing the Universal ISIS-PNP database, a user could discern the type of boiler used at the plant and, in turn,
Base Emissions (regular LNB)
Energy Efficiency Case (ULNB)
Pollutants
6-5
-------
infer flow mixing conditions in the boiler. For boilers with favorable mixing conditions, the user may analyze
application of SNCR in lieu of more expensive SCR. By comparing plants with boilers of different sizes, a user can
understand the economy of scale for wet FGD installation. Understanding the economy of scale may result in
installation of SDA rather than wet FGD for plants burning mid-sulfur fuels. Utilizing the Universal ISIS-PNP
database, the user can analyze reduction of emissions of other pollutants such as mercury since fuel properties
and installed air pollution control technologies are known at plant level.
Similarly, PM control technology applications could be analyzed to accomplish maximum PM emission reduction.
For example, a plant with excessive PM emissions despite having an ESP installed could have a fabric filter added
based on results from the Universal ISIS-PNP database that determined high resistivity PM (function of fuel use)
was limiting the performance of the ESP. This selective application of efficient air pollution control technologies
to plants with the highest emissions of a specific pollutant could then be analyzed by Universal ISIS-PNP to
understand how the cost impact of equipment installation could be minimized across the sector.
There is a menu of GFIG emission reduction measures for existing boilers (USEPA, 2010). Some of the most
commonly used measures include good O&M measures, air preheaters and economizers, boiler insulation,
minimization of inleakage, and steam line maintenance. The majority of measures are common, such as the
burner retrofit discussed above, yet capable of substantial CO2 emission reduction. In the example discussed
above, simple replacement/retrofit of burners was capable of approximately 6 percent CO2 emission reduction.
Other measures may be complex and may require site reconfiguration, such as, for example, combined heat and
power or repowering. Impacts of any measure are highly site-specific in terms of energy efficiency gain. In turn,
CO2 emission reduction corresponds to actual percent efficiency gain realized as the effect of measure
implementation. Using the Universal ISIS-PNP database and the menu of GFIG emission measures, the user is
able to optimize technology solutions that may be applied to specific boilers to reduce overall GFIG emissions
from the sector. As illustrated by the cases above, simultaneous reductions of GFIG and other pollutants may be
accomplished by measures such as fuel switching or energy efficiency improvements. Similar analyses can be
made for pollutants other than GFIG, utilizing a menu of SO2, NOx and other emission control technologies built
into Universal ISIS-PNP.
6.5. References
NCASI (2009). National Council for Air and Stream Improvement. Environmental Footprint Comparison Tool:
Trade-Offs and Co-Benefits Accompanying SOx and NOx Control. 2009.
USEPA (2010). US Environmental Protection Agency. Available and Emerging Technologies for Reducing
Greenhouse Gas Emissions from the Pulp and Paper Manufacturing Industry. Available at:
http://www.epa.gov/nsr/ghgdocs/pulpandpaper.pdf. Last accessed on October 31, 2014.
6-6
-------
The objective of this chapter is to demonstrate various aspects of running the Universal ISIS-PNP by introducing
the user to the model's user interface. The manual discusses the user's personal computer hardware/software
requirements for installing Universal ISIS-PNP. Usage instructions are given for opening and running the model.
Model input data requirements, data processing, and output are discussed. To familiarize user with the Universal
ISIS-PNP, examples of running different scenarios are given. Troubleshooting is also explained in the manual.
7.1. Hardware/Software Requirements of Universal ISIS-PNP
An overview of hardware requirements for the installation and simulation runs using the Universal ISIS-PNP
software package is given below in Table 7-1. The software package containing the database and executable files
can be downloaded from the USEPA website.
Table 7-1. System Requirements for Software Installation
Component
Requirements
Processor
i7-XXX (4th Generation)
Operating System
Windows 7 or Higher
Memory
Minimum 8 GB RAM
Support Software
GAMS-IDE
Microsoft Office 2007 or Higher
Latest Version of Adobe Acrobat Reader
Support Documents
GAMS Tutorial and User Manual: http://gams.com/docs/document.htm
7.2. Installation of GAMS (Supporting Software)
GAMS is a user interface to the UISIS PN P model that facilitates the running of the model, thereby allowing the
user to solve mathematical equations in the model to get optimized results. The aforementioned hardware and
software requirements of Universal ISIS-PNP apply to GAMS. The user should read the GAMS tutorial and user
manual document available on GAMS website (www.gams.com ) to get an overview of the installation process.
1.	Depending on the computer configuration (32-bit or 64-bit), select the appropriate GAMS executable
program file from the GAMS Website: http://www.gams.com/download/.
2.	Download the executable GAMS program file windows_x86_.32.exe (for 32-bit) or
windows_x86_64.exe (for 64-bit).
3.	Double click on the file to install and run the setup.
4.	Check the "Use advanced installation mode" box as shown in Figure 7-1.
7-1
-------
GAMS


Setup - GAMS
Welcome to the GAMS Setup
Wizard
This will install GAMS win64 24.0.1 on your computer,
It is recommended that you dose all other applications before
continuing.
Click Next to continue, or Cancel to exit Setup.
@ Use advanced instaBation mode
Next >
Cancel
Figure 7-1. GAMS Setup Wizard
5. Select a folder where GAMS wiil be installed. To carry out a Universal ISIS PNP model run, ensure that
GAMS is installed in the PATH that includes GAMS installation directory as shown in Figure 7-2.
Setup - GAMS
Select Destination Location
Where should GAMS be installed?
GAMS
Setup will install GAMS into the following folder.
To continue, dick Next. If you would like to select a different folder, dick Browse.
asaassssE
Browse..
At least 365.3 MB of free disk space is required.
< Back
Next >

Cancel
Figure 7-2. Select Destination Location
6. In "Advanced options," check all three boxes including "Add GAMS directory to PATH Environment
variable" as highlighted in Figure 7-3.
7-2
-------
Setup - GAMS
Advanced Options
Select advanced options
_ n
GAMS
0 Install GAMS for all users
0 Add GAMS directory to PATH environment variable
@ Create a desktop icon
< Back
Next >

Cancel
Figure 7-3. GAMS Setup
7. Select the gamslice.txt file and select "open" to perform copy. Copy the GAMS license file (gamslice.txt)
and choose the "Copy license file" option, as shown in Figure 7-4.
Setup - GAMS
_ n
GAMS


Completing the GAMS Setup
Wizard
Setup has finished installing GAMS on your computer. The
application may be launched by selecting the installed icons.
Click Finish to exit Setup.
0	Launch GAMS IDE
1	I Show release notes
GAMS license options:
O No license, demo only
® Copy license file
O Copy license text from clipboard
clipboard has no GAMS license
Finish
Figure 7-4. GAMS Setup
7.3. Opening and Running Universal ISIS-PNP
Download the Universal ISIS-PNP zipped file (containing the database and the executable files) from the USEPA
website. Folder contains:
7-3
-------
1)	Database (input and output Excel files)
2)	Universal ISIS-PNP (GAMS format)
3)	Model Interface
Note: Do not change the names of the Input and Output Microsoft Excel files because Universal ISIS-PNP GAMS
project file code will then become invalid and non-executable. It is also important to note that the Input and
Output Excel files should be stored in the same folder for GAMS to produce comprehensible results.
7.3.1. Open Project
To open a project, the steps are as follows:
1)	Open the GAMS-IDE program from the directory it has been installed in. For example, if the GAMS-IDE
program is installed in program files, click the following: Go to Start -> All Programs -> GAMS ->
GAMSIDE.
2)	To open Universal ISIS-PNP project, click on [File] at the top left of the GAMS toolbar and choose
[Project] -> [Open Project].
3)	The Universal ISIS-PNP project file will appear automatically on GAMS home screen and will have the
extension ".gpr" as shown in the Figure 7-5.
IDE
gamside: E:\2014 UISIS_PNP\UISIS_PROJECT.gpr
File Edit Search Windows Utilities Model Libraries Help

V
v r
~3
%
Figure 7-5. Universal ISIS-PNP Project Opens on GAMS
7.3.2. Open and Run Database
To open and run the database, the steps are as follows:
1) To open UISIS PNP database, click on the open folder shown circled in red in the figure above and choose
GAMS file DATABASE with extension *.gms from 2014 UISIS PNP folder, as shown in Figure 7-6.
7-4
-------
Open
Recent Places
Desktop
Libraries
Computer
4
Network
S3
2014 UISIS_PNP
J
Date illodtt*... Type
Name

DATABASE

6/3/20141...
Filefolder
lit. INPUTS

5/15/2014...
File folder
. MODEL

5,15/2014 ...
Filefolder
OUTPUTS

5./15/2014 ...
Filefolder
PROCESSING

5/15/2014 ...
Filefolder
- Cleanup

4/11/2013...
GAMS IDE file
fs] DATABASE

12/18/201...
GAMS IDE file
0 UISIS_PNP

12/18/201...
GAMS IDE file
|database
File name:
Files of type: | Gams files f.gms)
l~~ Open as read-only
U
Open
Cancel
Figure 7-6. Database GAMS File in Universal ISIS-PNP Folder
2)	Once the user clicks on the DATABASE file, the file opens with title bar reading [s=Database], as shown
in Figure 11, indicating that the database will be saved by GAMS and will get ready to be processed. An
empty title bar indicates that the database will not be saved for the model to process it. To avoid this
issue, user can manually type s=Database in the empty space.
e* pass	3
Figure 7-7, GAMS Title Window before Running Database
%
3)	Click on the [Run] button to run the database. Once user hits run, a new window titled "1 active
process" should open. This signifies that GAMS is actively running on code file.
4)	The active process window indicates the status of the program's run and if any troubleshooting is
required. When the status reads [Normal Completion], as shown in Figure 7-8, the GAMS file has run
successfully.
Note: GAMS generates a new window for database.1st file.The Extension "1st" stands for Data list file. One can
expand the "display" button in database.1st to see a list of input parameters and variables.
7-5
-------
m	No active process	! I 0 ItfM
database
—
.Cal Imp&Exports. gms(80) 5 Mb a
—
DATABASE.gms(46) 5 Mb
—
.Cal Controls. gms(214) 5 Mb
—
DATABASE . gms (47) 5 Mb
—
.Cal Policy.gms(179) 5 Mb
—
DATABASE. gms(48) 5 Mb
—
.Cal EscRates.gms(14) 5 Mb
—
call GDXXRW.EXE DATABASE\UISIS_PNPINPUTS.XLSX @INPUTS\PnPInputs.txt
GDXXRW Dec 13, 2012 24.0.1 WIN 37366.37409 VS8 X86/MS Windows
Input file : D:\ACE 139 - UISIS PnP FROJECT\UI5IS GAMS DEVELOPMENTS014 U3
Output file: D:\ACE 139 - UISIS PnP PROJECT\UISIS GAMS DEVELOPMENTS014 TJ1
Total time = 2062 Ms
	
.Cal EscRates.gms(16) 5 Mb
	
GDXin=D:\ACE 139 - UISIS PnP PROJECT\UISIS GAMS DEVELOPMENTS014 UISI5
	
.Cal EscRates.gms(474) 5 Mb
	
DATABASE.gms(56) 5 Mb
	
Starting execution: elapsed 0:00:31.669
	
DATABASE.gms(2302) 18Mb
	
Putfile dfx D:\ACE 139 - UISIS PnP PROJECT\UISIS GAMS DEVELOPMENTS01^
* * #
Status: Normal completion
	
Job DATABASE.gms Stop 06/05/14 13:35:46 elapsed 0:00:32.041
<
>

Figure 7-8. Model Run Status
7.3.3. Open and Run Model
1) Clickon [OPEN] folder icon to choose IJISIS PNP GAMS file from t he IJ ISIS PNP folder, as shown
in Figure 7-9.
a Open
Look in:
%
Recent Places
Desktop
3
Libraries
iMs:
Computer
%
Network
S3 J
| , 2014 UISIS_PNP
Name
M DATABASE
J INPUTS
jl MODEL
^ OUTPUTS
J PROCESSING
1'gj Cleanup
® DATABASE
-iUISIS PNP
< I	
File name:
Files of type:
[uisis.p
I Gams files f.gms)
Open as read-only
"3
«- B Eh
Date modif... Type
6/3/20141... File folder
5/15/2014 ..
5/15/2014 ..
5/15/2014..
5/15/2014 ..
4A1/2013..
12/18/201...
12/18/201...
1]
T3
File folder
File folder
File folder
File folder
GAMS IDE file
GAMS IDE file
GAMS IDE file
Open
Cancel
Figure 7-9. Choosing Universal ISIS-PNP GAMS File
7-6
-------
2) The file opens with the title bar reading [r=Database], as shown in Figure 7-10.
F
r=Database
U
Figure 7-10: GAMS title window before running the model
3)	This means the database that was saved earlier is ready to be run through the mathematical
equations of the model. An empty title bar indicates that the database will not run. To avoid
this issue, user can manually type r=Database in the empty space.
4)	At this point, both the model and database in GAMS format are displayed next to each other
as shown in Figure 7-11.
££ gamside: D:\ACE 139 - UISIS PnP PROJECT\UISIS GAMS DEVELOPMENT\2014 UISIS.P...
File Edit Search Windows Utilities Model Libraries Help
3 _w]_#]jv] l^'UH+W
"3 m
.. D:\ACE 139 - UISIS PnP PROJECt\UISIS GAMS DEVELOPMENT\2014 UISIS_PNP\D... r^T~ir|£sJ
DAT ABAS E.gms UISIS_PNP.gms |
SOfftext
$echo "1 Compilation Error' > Outputs\status.txt
file dfx / "Outputs\i
$Title Universal Industrial Sector Integrated Solution Module For Pulp And P
SOnUNDF
$Offsymlist Offsymxref
*Eguation listing (limrow) and variables listing (limcol)
limcol = 0, solprint = off;
Option solveopt = replace;
Option limrow =
Universal Industrial Sector Integrated Solution Module
For Pulp and Paper Sector
by
APTB/APPCD/NRMRL
United States Environmental Protection Agency
109 T.B. Alexander Drive, Raleigh, NC 27711
Email: bhander.gurbakhash@epa.gov
. txt" /; put dfx "2 Execution Error"; putclose dfx;
Figure 7-11: Database and Universal ISIS-PNP GAMS files
%
5)	After making sure r=Database shows up on the screen, click on [Run]	I button.
6)	Check the status of completion at the bottom of the window. A successful UISIS_PNP model
run should display 'Normal Execution' status and the numbers for [MIP Solution], [Final Solve]
and [Best possible] should match as shown in Figure 7-12.
7)	Take note of location of the output file marked in yellow in the figure below. All results are
displayed in the excel sheet at this location.
7-7
-------
No active process	l-cj-lfa
database uisis_pnp
MIP
Solution: -731372892.
664440 (21899 iterations.
0 nodes)
A
Final Solve: -731372892.
664437 (16094 iterations)







Best possible: -731372892.
664440


Absolute gap: 0.
000000


Relative gap: 0.
000000


	
Restarting execution



	
UISIS PNP.gms(4483) 611
Mb


—
Reading solution for model UISIS ElPnP


***
Reading with solveopt=REPLACE (0)


	
UISIS_PNP.gms(4483) 612
Mb


	
Executing after solve:
elapsed 0:01:18.222


	
UISIS_PNP.gms(4484) 613
Mb


	
UISIS_PNP.gms(5066) 619
Mb


GDXXRW Dec 18, 2012 24.0.1 WIN 37366.37409 VS8 X86/MS Windows

Input file : D:\ACE 139 - UISIS PnP PROJECT\UISIS GAMS
DEVELOPMENT\2 014
U]
Output file: D:\ACE 139 - UISIS PnP PROJECT\UISIS GAMS
DEVE LOPMENT\2 014
U]
Total time = 23140 Ms



	
UISIS_PNP.gms(5130) 625
Mb

V
<



>

Figure 7-12. Universal ISIS-PNP Output File Location
7.4. Model Input Requirements
The database was organized in Microsoft SQL database and an Excel Workbook was developed based on
information derived from the 2002 U.S Economic Census of pulp, paper and paperboard mills (North American
Industry Classification system (NAICS) Code 3221) and data obtained from RISI and boiler MACT. The interface
allows the user to enter the data, which include historical and projected nationwide commodity consumption,
imports, exports, number of production facilities, distance from production facilities to the demand centers,
production capacity, associated costs (e.g., material, operations, and maintenance), fuel types and costs,
emissions sources and intensities, and other data.
Various tabs in the input Excel sheet of this data set are explained in the table below to familiarize the user with
location of desirable inputs.
POLICY
CONSUMPTION
PLUNITS
PRUNITS
Policy parameters for reduction of emissions such as, for example, target year,
percentage emission reduction, etc., can be defined here
Product demand in all three regions (North, South, and West) from 2007 to 2020
Unit level data for integrated and non-integrated pulp mills. Information includes
plant location, cost of raw materials, labor, repair and maintenance, total electricity
produced and consumed, black liquor and lime mud quantity produced by mill.
Production capacity, types of fuel and boiler availability by production unit.
Unit level data for integrated and non-integrated paper mills. Information includes
plant location, cost of raw materials, labor, repair and maintenance and total
electricity produced and consumed by product in each mill. Production capacity,
hardwood, softwood, recycling, additive pulp consumption percentage by product
and types of fuel and boiler availability by production unit.
7-8
-------
BLRSCAP
RMT&RECYCLING
ENERGY
ENGINTENSITY
INTENSITY
TRANS&IMPORTS
EXPORTS
ESCALATION RATES
WASTEWATER
CONTROLS
PLCONTROLS
PRCONTROLS
CALIBRATION
Boiler data for pulp and paper mills. The data include boiler capacity, type of fuel
burned, type of control installed on the boiler.
Recycling and Transportation costs from demand centers to supply centers
Fuel cost for North, South, and West regions. The data also include 2011 electricity
price by State.
Fuel/Energy data by mill
Emissions data by fuel and heat intensity of individual fuels
Product transportation costs, import quantity and price by North, West, and South
regions
Product transportation costs, export quantity and price by North, West, and South
regions
Escalation rates for all costs used in the model
Wastewater and landfill quantity and costs by mill
Control availability, capacity, capital costs and compatibility with boilers and fuels
Control data for pulp mills
Control data for paper mills
Model Calibration adjustments
7.5. Pre-Processing of Data
1)	Input data for the Universal ISIS-PNP are organized in a Microsoft Excel spreadsheet and in a Microsoft
SQL database.
2)	This user-defined input spreadsheet includes the time horizon (simulation period), reference year,
discount rate, time blocks, commodity characteristics, emissions types, fuel types, and plant types.
3)	GAMS communicates with this input data Excel sheet via GDX (GAMS Data Exchange) files. The
mathematical modeling framework coded into GAMS optimizes the input data for the optimal levels of
production, imports, and controls required to meet the demand.
4)	The optimized input data are exported to Microsoft SQL database (through Excel spreadsheets) for
further analyses.
Data processing is shown schematically in Figure 7-13, below.
7-9
-------


INPUTS
Historical and
Projected Price Data
MARKET
Emission Reduction,
Ta rget Taxes,
Elasticity, Time Frame
POLICY
Products, Capacity,
Fuel, and Energy
Efficiency.
INDUSTRY
o

Universal ISIS - Pulp
and Paper
GAMS-IDE Model


J

OUTPUTS

Energy
Efficiencies and
Emissions
(Controlled and
Uncontrolled)
Optimized Costs
of Commodities
Optimized
Mitigation
Options

J
Figure 7-13. Universal ISIS-PNP Data Structure
7.6. Output Database
The desirable outputs which include optimized production level of each production unit by product, imports in
each region, fuel requirements, controls, boiler emissions, and costs may be selected by the user. The resulting
computations are exported to the PnPOUTPUTS.gdx file or into the Excel file, UISIS_PNPOUTPUT.xlsx stored in
the same folder as the input file. Various tabs in the output Excel sheet of this data set are explained in the table
below to familiarize the user with location of desirable outputs.
AGGREGATE RESULTS Optimized quantity and costs of total production and imports for both pulp and
paper products for all mills combined.
PRICES
COSTS
PLPPRODN
FUELS
PRPPRODN
DOMTRANSPORT
Average and Marginal prices for all paper products
Fuel, raw material, labor, operation and maintenance, domestic transportation,
and annual costs of production
Optimized quantity and costs of pulp products by mill
Optimized quantity and costs of fuel by mill for both pulp and paper mills.
Optimized quantity and costs of paper products by mill
Optimized domestic transport quantity and costs by product for both pulp and
paper mills
7-10
-------
P LP	BO ILE RS
PR_BOILERS
BLS&MUD
PRIMP&EXPORTS
EMISSIONS
RECYCLE
Total heat and fuel required by mill for pulp mills
Total heat and fuel required by mill for paper mills
Black liquor solids and lime mud quantity by mill for pulp mills
Annual imports and export quantity by product
Annual domestic emissions by mill for both pulp and paper mills
Amount of recycled material used
7.7. Running a Scenario in Universal ISIS-PNP
The model evaluates environmental and economic impacts of emission reduction scenarios by comparing the
scenario case with business as usual case. Three cases are presented in this manual to make the user conversant
with manipulating input data to obtain desired results from the model:
•	BAU
•	Scenario I - Emission constraints
•	Scenario II - Fuel constraints
7.7.1. Business as Usual (BAU)
Step 1: Open Universal ISIS-PNP project (See section 7.3.1)
Step 2: Open UISIS PNPINPUTS excel sheet in DATABASE sub-folder of 2014 Universal ISIS-PNP folder, as shown
in Figure 7-14.
@
Vlfl ~ Computer ~ DATA (E:) ~ 2014 UISIS_PNP ~ DATABASE
Search DATABASE
Organize ~ Include in library ~ Share with
New folder
i== ^ si
"iV Favorites
H Desktop
^ Downloads
1i Recent Places
Name
PL UISIS_PNPINPUTS
H UISIS_PNPOUTPUTS
Date modified
6/2/20143:17 PM
5A5/2014 4:20 PM
Type
Microsoft Excel W„,
Microsoft Excel W,.,
Size
1,295 KB
3,607 KB
Figure 7-14. Input and Output Database Files of Universal ISIS-PNP
Step 3: Go to GENERALS tab to define the time period of simulation under General Inputs (Pulp and paper mills).
The cost data in the model were populated for the reference year 2009, and the simulation starts in 2010. To
change the reference year to something other than 2009, the data in the input sheet have to be updated to
reflect costs for that year. The user can choose to change the last year of simulation (e.g., lastsimyear, as shown
in Figure 7-15). The optimized results will be projected until this last year. For this example, 2020 is chosen as
the target year. In this case, the model will run simulations for a ten-year period (2010-2020).
7-11
-------


Time

simrefyear
2009

"Istsimyear
2010

lastsimyear
2020

time block
1



Figure 7-15: Choosing time period of simulation in input file
Step 4: Go to POLICY tab of the spreadsheet to manipulate policy parameters for reduction of emissions (see
section 7.4). For BAU scenario, scenario parameters should not be applied ("0" in Figure 7-16). Save and close
the Excel file.
Allowances Auctioned at Fixed Allowance Price
1
Allowances Grandfathered
0
Allowance Trading with Output-Based Rebating
0
Apply Policy Parameters (Important: Apply Policy = 1, BAU = 0)
0
Figure 8-16: Applying business as usual parameters in input file
Step 5: Open and run DATABASE (See section 4.2).
Step 6: Open and run the model (See section 4.3).
Step 7: Open Universal ISIS-PNPOUTPUTS excel sheet in DATABASE sub-folder of 2014 Universal ISIS-PNP folder.
The results are displayed in this output file. See Section 7 for location of desirable outputs.
7.7.2. Scenario I - Emission Constraints
Step 1: Open Universal ISIS-PNP project (See section 4.1).
Step 2: Open Universal ISIS-PNPINPUTS excel sheet in DATABASE sub-folder of 2014 Universal ISIS-PNP folder.
Go to GENERALS tab to define the time period of simulation under General Inputs (Pulp and paper mills). See
Step 3 of Section 7.7.1.
Step 3: Go to POLICY tab of the spreadsheet to change to "Apply Policy Parameters" ("1" in Figure 7-17).
7-12
-------
Allowances Auctioned at Fixed Allowance Price
1
Allowances Grandfathered
0
Allowance Trading with Output-Based Rebating
0
Apply Policy Parameters (Important: Apply Policy = 1, BAU = 0)
>¦1
Figure 7-17. Applying Policy Parameters in Input File
Step 4: Set emission reduction constraints on POLICY tab under EMISSION INTENSITY TARGETS for target year,
as shown in Figure 7-18. For example, Pollutants in consideration column gives values of NOx pollutant in BAU
scenario between 2010 and 2020. To achieve 50-percent reduction of NOx by target year 2020, value of NOx in
Emissions Intensity Targets column should be half of its 2010 value in Pollutants in consideration column. User
can choose to change the target year to the year by which you want to meet the emission reductions. Save and
close the UISIS PNPINPUTS Excel file after having specified the policy parameters.
Target Year
2020

Emissions Intensity Targets
EMISSION
Value
NOx
=0.5*44468
CO 2

SO 2

PM

TRS

voc.

HAP



Pollutants in Consideration (S.Tons/year)

NOx
CO 2
S02
2010
44,468
12,387,633
77,017
Figure 7-18. Target Emissions Constraint in Input File
Step 5: Open and run DATABASE (See section 4.2).
Step 6: Open and run the model (See section 4.3).
Step 7: Open UISIS_PNPOUTPUTS excel sheet in DATABASE sub-folder of 2014 UISIS PNP folder. The results are
displayed in this output file. See section 7 for location of desirable outputs.
7.7.3. Scenario II- Fuel Constraints
Step 1: Open Universal ISIS-PNP project (See section 4.1).
Step 2: Follow Step 2 and 3 of Section 8.2.
7-13
-------
Step 3: Set fuel constraints on POLICY tab under Pulp Mills Fuel Constraints and Paper Mills Fuel Constraints
columns for fuels, as shown in Figure 7-19. For example: - To make the model satisfy market demand without
use of coal, set the value of 0.000001 for coal under "Fuel Constraints (=>) Note: Cheap fuel constraints must be
less than orequal to" column for both pulp and paper mills fuel constraints. Save and close the UISIS PNPINPUTS
Excel file after you have specified the policy parameters.
Note: If user wants the model not to utilize coal, user cannot enter "0" value for coal because the model is coded
to read "0" as an empty box/no value. Entering a negligible value like 0.0000001 is understood by the model as
constraint and it limits the use of coal. Entering "0" will make the model fully ignore the constraint resulting in
the same results as BAU scenario.
Pulp Mills Fuel Constraints
Coal
NG
Oil
Hog
Fuel Constraints (<=) Note: Expensive fuel constraints must be greater than or equal to
Fuel Constraints (=>) Note: Cheap fuel constraints must be less than or equal to
0.OOOOOO1
Coal
NG
Oil
Hoe
Paper Mills Fuel Constraints
Coal
NG
Hog
Fuel Constraints (<=) Note: Expensive fuel constraints must be greater than or equal to
Fuel Constraints (=>) Note: Cheap fuel constraints must be less than or equal to
O OOOOOOl
Coal
NG
Oil
Hog
Figure 7-19. Fuel Constraints in Input File for both Pulp and Paper Mills
Step 4: Open and run DATABASE (See section 4.2).
Step 5: Open and run the model (See section 4.3).
Step 6: Open UISIS_PNPOUTPUTS Excel sheet in DATABASE sub-folder of 2014 Universal ISIS-PNP folder. The
results are displayed in this output file. See section 7.6 for location of desirable outputs.
7.8. Troubleshooting
GAMS error messages can be broadly classified into two types - compilation errors and execution errors.
7.8.1. Compilation Errors
The most common errors encountered when using GAMS-IDE are the compilation errors. Compilation errors
arise due to errors in the syntax of the GAMS code. Common compilation errors include:
• Forgetting to end a statement or operation line with a
7-14
-------
•	Misspelled words or commands
•	Using a parameter or variable before defining the parameter or variable.
Compilation error messages are outlined in red in the active process window. Click on these messages to be
navigated to where the error was noticed. The Data list file contains markers for compilation errors as well. A
text set marked by "****" in the beginning of the line signals compilation errors. Error messages at the end of
the Data list file provide more detail on these errors.
7.8.2. Execution Errors
Execution errors are more complex and difficult to decipher than compilation errors. Execution errors may
include incorrect equation specification, insufficient memory, or infeasibility error.
Incorrect equation specification
Execution errors often arise from incorrectly specified equations and models or unrecoverable constraints such
as dividing by zero. Expanding on display statements and checking on defined parameters and variables to check
on output values may prove useful in trying to single out the sources of executable errors.
Insufficient memory
Another common execution error is caused by insufficient memory. When this error is encountered the, user
should try to install GAMS on a machine with higher memory (4 GB RAM (64-bit)). Alternatively, the user may
choose to use an extended virtual memory to solve this problem although it will slow down the GAMS program.
The insufficient memory error display is shown in Figure 7-20.
¦« 1 active process 1 <=> 1 ®

uisis_pnp |
Iteration: 13097 Dual objective = -1063112914.612150
Elapsed time = 327.93 sec. (381244.74 ticks, 15051 iterations)
Elapsed time = 337.55 sec. (392603.98 ticks, 15288 iterations)
Iteration: 15975 Dual objective = -1048174272.772749
Elapsed time = 346.10 sec. (402639.86 ticks, 16318 iterations)
Elapsed time = 354.28 sec. (412775.01 ticks, 16708 iterations)
Elapsed time = 362.62 sec. (422936.69 ticks, 17403 iterations)
Elapsed time = 371.14 sec. (433298.19 ticks, 17633 iterations)
Iteration: 18609 Dual objective = -1010150231.220840
Elapsed time = 379.21 sec. (443376.93 ticks, 18828 iterations)
Elapsed time = 392.56 sec. (460021.76 ticks, 18872 iterations)
Removing perturbation.
Root relaxation solution time = 400.36 sec. (469234.56 ticks)
Root node processing (before b£c):
Real time = 632.43 sec. (522275.29 ticks)
Sequential b&c:
Real time = 0.00 sec. (0.00 ticks)
Total (root+branchscut) = 632.43 sec. (522275.29 ticks)
MIP status(107): time limit exceeded
Cplex Time: 1007.64sec (det. 555849.14 ticks)
Fixing integer variables, and solving final LP...
Presolve time = 1.44 sec. (12.22 ticks)
Insufficient memory for presolve.
CPLEX Error 1001: Out of memory.
Fixed MIP status(0):
Cplex Time: 3.25sec (det. 655.02 ticks)
«»« CPLEX Error 1001-r"Out o£ memoryT^)

4 1 ,1 ~
«** CPLEX Error 1001: Out of memory.
Interrupt | Stop J f~ Summary only W Update
Figure 7-20. Insufficient Memory Error in Universal ISIS-PNP Run
Infeasibility Error
The Universal ISIS-PNP utilizes several mathematical equations to satisfy a known demand at least cost while
applying emission constraint inputs by the user. When the applied user constraints curb the ability of the model
7-15
-------
to balance both the left and right hand side of the equations, the model fails to execute normal completion and
displays an infeasible solution error. For example: The user puts a restriction on coal to get the demand met by
other fuels like natural gas, etc., but the model requires energy from coal along with other fuels to satisfy heat
balance equations. The model runs in two parts, general and elastic run. Even if the running status shows normal
completion at the end of the Universal ISIS-PNP GAMS window, the user must check individual solutions of both
models to check infeasibility error in each of them. The infeasibility error display is shown in Figure 7-21.
¦ ¦ No active process
database uisis_pnp
a ss
Starting Cplex...
Row 'EQFRDEMAND(2010.North.CNT)1 infeasible, all entries at implied bound3.
Presolve time = 1.44 sec. (680.91 ticks)
MIF status(103): integer infeasible	
Cplex Time: 1.56sec (det. 772.50 ticks)
CFLEX Error 1217: No solution exists.
Problem is integer infeasible.
-- Restarting execution
¦- UISIS_PNP.gras(4392) 167 Mb
-- Reading solution for model UI5IS PnP
GENERAL MODEL
SOLUTION
*** Reading with solveopt=REPLACE
* * *
(0)
Executing after solve: elapsed 0:00:24.473
OISIS_FNF.gms(43 94) 163 Mb
UISIS_PNP.gms(4395) 163 Mb
Et-TF.oir.B (44551—1£8 Mb	
SDXXRW
Dec 18. 2012 24.0.1 WIN 37366.37409 V58 X36/M5 Windows
XL5X
Input file : E:\2014 UISIS_PNP\EMISSIONS.GDX
Output file: E:\2014 UI5I5_PNP\DATA3A5E\UISI5_FNFINFUT5
total time = 2309 Ms
	 UI5I5_FNF.gms(4432) 163 Mb
	 Generating MIF model UISIS_ElPnP
	 DISIS_PNP.gras(4433) 681 Mb
1,530,141 rows 3,160,488 columns 7,861,007 non-
1,225,334 discrete-columns
	 Executing CPLEX: elapsed 0:00:38.410
IBM ILOG CFLEX Dec 18, 2012 24.0.1 WEX 37366.37409 WE:
	 GAMS/Cplex licensed for continuous and discrete problems.
Cplex 12.5.0.0
Reading data...
Starting Cplex...
Row 1EQFRELDEMAND(2010.North.CNT) " infeasible, all entr:
Presolve time = 1.36 sec. (1082.41 ticks)
MIF status(103): integer infeasible
Cplex Time: 1.96sec (det. 1174.77 ticks)
CFLEX Error 1217; No solution exists.	
ELASTIC MODEL
SOLUTION
es at implied bounds.
Problem is integer infeasible.
	 Restarting execution
Figure 7-21. Infeasibility Error in Universal ISIS-PNP Run
7-16
-------
For more information on troubleshooting, licensing questions or other general queries, please refer to the GAMS
tutorial and User Guide - http://www.gams.com/docs/document.htm
7-17
-------
Appendices
Appendix A RTI International Draft Memo for Pulp and Paper Industry January 16, 2009
Appendix B Andover Technology Partners Memo March 15, 2010
Appendix C RTI International Memos for Paper Machine November 16, 2011 and March 29, 2013
1
-------