United Taconite Analysis of Best Available Retrofit Technology (BART)

Transcription

1 United Taconite Analysis of Best Available Retrofit Technology (BART) Table of Contents 1. Executive Summary...iv 2. Introduction A BART Eligibility B BART Determinations Streamlined BART Analysis A Indurating Furnaces B PM-Only Taconite MACT Emission Units C Sources of fugitive PM that are subject to MACT standards D Non-MACT Units and Fugitive Sources (PM only) E Other Combustion Units F Visibility Impact Modeling for Negligible Impacts Baseline Conditions and Visibility Impacts for BART Eligible Units A MPCA Subject-to-BART Modeling B Facility Baseline Modeling C Facility Baseline Modeling Results BART Analysis for Indurating Furnaces A Indurating Furnace A.i Sulfur Dioxide Controls A.i.a STEP 1 Identify All Available Retrofit Control Technologies A.i.b STEP 2 Eliminate Technically Infeasible Options A.i.c STEP 3 Evaluate Control Effectiveness of Remaining Control Technologies A.i.d STEP 4 Evaluate Impacts and Document the Results A.i.e STEP 5 Evaluate Visibility Impacts A.ii Nitrogen Oxide Controls A.ii.a STEP 1 Identify All Available Retrofit Control Technologies A.ii.b STEP 2 Eliminate Technically Infeasible Options A.ii.c STEP 3 Evaluate Control Effectiveness of Remaining Control Technologies A.ii.d STEP 4 Evaluate Impacts and Document the Results A.ii.e STEP 5 Evaluate Visibility Impacts Visibility Impacts A Post-Control Modeling Scenarios B Post-Control Modeling Results Select BART...68 i

2 List of Tables Table 3-1 Summary Table of BART-Eligible Units Subject to a Streamlined BART Analysis Table 4-1 Baseline Conditions Modeling Input Data and the Basis for 24-hour Emissions Data. 17 Table 4-2 Baseline Visibility Modeling Results Table 5-1 Indurating Furnace SO 2 Control Technology Availability, Applicability, and Technical Feasibility Table 5-2 Indurating Furnace SO 2 Control Technology Effectiveness Table 5-3 Indurating Furnace SO 2 Control Cost Summary Table 5-4 SO 2 Control Technology Impacts Assessment Table 5-5 Post-Control Modeling Scenarios Table 5-6 Post-Control SO 2 Modeling Scenarios - Modeling Input Data Table 5-8 Post-Control SO 2 Modeling Scenarios - Visibility Modeling Results Table 5-9 Post-Control SO 2 Modeling Scenarios - Visibility Improvements Table 5-10 Indurating Furnace NO x Control Technology Availability, Applicability, and Technical Feasibility Table 5-11 Indurating Furnace NO x Control Technology Effectiveness Table 5-12 NO x Control Cost Summary Table 5-13 NO x Control Technology Impacts Assessment Table 5-14 Post-Control Modeling Scenarios Table 5-15 Post-Control NO X Modeling Scenarios Emissions Data Table 5-16 Post-Control NO X Modeling Scenarios - Visibility Modeling Input Data Table 5-17 Post-Control NO X Modeling Scenarios - Visibility Modeling Results Table 5-18 Post-Control NO X Modeling Scenarios - Visibility Improvements Table 6-1 Post-Control Modeling Scenarios Table 6-2 Post-Control NO X Modeling Scenarios - Modeling Input Data Table 6-3 Post-Control Modeling Scenarios - Visibility Modeling Results Table 6-4 Post-Control Modeling Scenarios - Visibility Modeling Results Table 6-5 Post-Control Modeling Results Dollars per Deciview Improvement List of Figures Figure 2-1 Minnesota s BART Geography... 2 ii

3 List of Appendices Appendix A Appendix B Appendix C Appendix D Appendix E Control Cost Analysis Spreadsheets Changes to MPCA BART Modeling Protocol Visibility Impacts Modeling Report Applicable and Available Retrofit Technologies Clean Air Interstate Rule (CAIR), Cost-Effective Air Pollution Controls iii

4 1. Executive Summary United Taconite is located in northern Minnesota, with mining facilities near Eveleth, Minnesota. United Taconite s processing facilities are located near Forbes, Minnesota. This report describes the background and methods for the selection of the Best Available Retrofit Technology (BART) as proposed by United Taconite for its taconite processing plant located near Forbes, MN. MPCA identified approximately 70 pieces of equipment at United Taconite that were installed within the time window ( ) that makes them subject to BART. The equipment includes two gratekiln indurating furnace lines. Line 1 is permitted for natural gas and fuel oil, but Line 2 also uses coal and petroleum coke. Other equipment identified includes crushers, pellet coolers, combustion units, fugitive sources, and material handling and/or storage units for ore, product, and additives. Preliminary visibility modeling conducted by the Minnesota Pollution Control Agency (MPCA) found that air emissions from United Taconite s facility cause or contribute to visibility impairment in a federally protected Class I area, therefore making the facility subject to BART. Guidelines included in 40 CFR 51 Appendix Y and MPCA Attachments 2 and 3 were used to propose BART for the subject units. The existing pollution control equipment includes a wet scrubber designed to control particulate matter (PM); however some collateral control of sulfur dioxide (SO 2 ) is also achieved. A dispersion modeling sequence of CALMET, CALPUFF, and CALPOST was used to assess the visibility impacts of the baseline emissions and after the application of candidate BART controls. Visibility impacts were evaluated in the selection of BART. Other criteria that the BART rules require to be considered include the availability of technology, costs of compliance, energy and environmental impacts of compliance, existing pollution control technology in use at the source, and the remaining useful life of the source. Based on consideration of all of the above criteria, United Taconite proposes the following as BART: SO 2 emissions will be controlled by the existing scrubbers. NO x emissions will be controlled by the implementation of a heat recuperation project on Line 1. iv

5 PM emissions will be controlled as prescribed by the taconite maximum achievable control technology (MACT) standard 1. CALPUFF model is conservative, resulting in an over prediction of impacts. This modeled high impact from the BART eligible sources is 4.7 dv, which is slightly above perceptible levels of one to two dv. Real impacts to the Class I areas from United Taconite are expected to be even less than these modeled impacts. The proposed BART controls will reduce the facility s visibility impact by 0.7 dv, an improvement of 15%, on the worst visibility days. This represents a significant contribution to the default glide path of reducing impacts by 1-2% per year CFR Subpart 63 RRRRR-NESHAPS: Taconite Iron Ore Processing v

6 2. Introduction United Taconite is located in northern Minnesota, with mining facilities near Eveleth, Minnesota. United Taconite s processing facilities are located near Forbes, Minnesota. This report describes the background and methods for the selection of the Best Available Retrofit Technology (BART) as proposed by United Taconite for its taconite processing plant located near Forbes, MN. To meet the Clean Air Act s requirements, the U.S. Environmental Protection Agency (U.S. EPA) published regulations to address visibility impairment in our nation s largest national parks and wilderness ( Class I ) areas in July This rule is commonly known as the Regional Haze Rule [64 Fed. Reg (July, 1999) and 70 Fed. Reg (July 6, 2005)] and is found in 40 CFR part 51, in through Within its boundary, Minnesota has two Class I areas the Boundary Waters Canoe Area Wilderness and Voyageurs National Park. In addition, emissions from Minnesota may contribute to visibility impairment in other states Class I areas, such as Michigan s Isle Royale National Park and Seney Wilderness Area. By December 2007, MPCA must submit to U.S. EPA a Regional Haze State Implementation Plan (SIP) that identifies sources that cause or contribute to visibility impairment in these areas. The Regional Haze SIP must also include a demonstration of reasonable progress toward reaching the 2018 visibility goal for each of the state s Class I areas. One of the provisions of the Regional Haze Rule is that certain large stationary sources that were put in place between 1962 and 1977 must conduct a Best Available Retrofit Technology (BART) analysis. The purpose of the BART analysis is to analyze available retrofit control technologies to determine if a technology should be installed to improve visibility in Class I areas. The chosen technology is referred to as the BART controls, or simply BART. The SIP must require BART on all BART-eligible sources and mandate a plan to achieve natural background visibility by Figure 2-1 illustrates the BART-eligible facilities and the two Class I areas in Minnesota. When reviewing Figure 2-1, it is important to note that Minnesota Steel Industries (MSI) and Mesabi Nugget (Nugget), which are illustrated in the figure, are not currently in operation. The SIP must also include milestones for establishing reasonable progress towards the visibility improvement goals and plans for the first five-year period. Upon submission of the Regional Haze SIP, states must make the requirements for BART sources enforceable through rules, administrative orders or Title V permit amendments. 1

7 Figure 2-1 Minnesota s BART Geography The Minnesota SIP will address the 2 PSD Class I Areas [Voyageurs National Park (VPN) and Boundary Waters Canoe Area (BWCA)] and BART-eligible units illustrated above. (Source MPCA BART-Strategy October 4, 2005) By U.S. EPA s definition, reasonable progress means that there is no degradation of the 20 bestvisibility days, and the 20 worst-visibility days must have no more visibility impairment than the 20 worst days under natural conditions by Assuming a uniform rate of progress, the default glide path would require 1 to 2 percent improvement per year in visibility on the 20 worst days. The state must submit progress reports every five years to establish their advancement toward the Class I area natural visibility backgrounds. If a state feels it may be unable to adopt the default glide path, a slower rate of improvement may be proposed on the basis of cost or time required for compliance and non-air quality impacts. Note that the improvements required under the Regional Haze regulations are different from the BART requirements. Facilities subject to BART are not required to make all of the reasonable 2 70 FR No. 178 pp to

8 progress towards improving regional haze in Class I areas. Rather, BART is but one of many measures which states may rely upon in making reasonable progress towards regional haze improvement goals. 2.A BART Eligibility BART eligibility is established on the basis on three criteria. In order to be BART-eligible, sources must meet the following three conditions: 1. Contain emission units in one or more of the 26 listed source categories under the PSD rules (e.g., taconite ore processing plants, fossil-fuel-fired steam electric plants larger than 250 MMBtu/hr, fossil-fuel boilers larger than 250 MMBtu/hr, petroleum refineries, coal cleaning plants, sulfur recovery plants, etc.); 2. Were in existence on August 7, 1977, but were not in operation before August 7, 1962; 3. Have total potential emissions greater than 250 tons per year for at least one visibilityimpairing pollutant from the emission units meeting the two criteria above. Under the BART rules, large sources that have previously installed pollution-control equipment required under another standard (e.g., MACT, NSPS and BACT) will be required to conduct visibility analyses. Installation of additional controls may be required to further reduce emissions of visibility impairing pollutants such as PM, PM 10, PM 2.5, SO 2, NO x, and possibly Volatile Organic Compounds (VOCs) and ammonia. Sources built before the implementation of the Clean Air Act (CAA), which had previously been grandfathered, may also have to conduct such analyses and possibly install controls, even though they have been exempted to date from any other CAA requirements. Once BART eligibility is determined, a source must then determine if it is subject to BART. A source is subject to BART if emissions cause or contribute to visibility impairment at any Class I area. Visibility modeling conducted with CALPUFF or another U.S. EPA -approved visibility model is necessary to make a definitive visibility impairment determination (>0.5 deciviews). Sources that do not cause or contribute to visibility impairment are exempt from BART requirements, even if they are BART-eligible. 2.B BART Determinations Each source that is subject to BART must determine BART on a case-by-case basis. Even if a source was previously part of a group BART determination, individual BART determinations must be made 3

9 for each source. The BART analysis takes into account six criteria and is analyzed using five steps. The six criteria that comprise the engineering analysis include: the availability of the control technology, existing controls at a facility, the cost of compliance, the remaining useful life of a source, the energy and non-air quality environmental impacts of the technology, and the visibility impacts. 3 The five steps of a BART analysis are: Step 1 - Identify all Available Retrofit Control Technologies The first step in the analysis is to identify all retrofit control technologies which are generally available for each applicable emission unit. Available retrofit control technologies are defined by U.S. EPA in Appendix Y to Part 51 (Guidelines for BART Determinations under the Regional Haze Rule) as follows: Available retrofit technologies are those air pollution control technologies with a practical potential for application to the emissions unit and the regulated pollutant under evaluation. Air pollution control technologies can include a wide variety of available methods, systems, and techniques for control of the affected pollutant. Technologies required as BACT or LAER are available for BART purposes and must be included as control alternatives. The control alternatives can include not only existing controls for the source category in question, but also take into account technology transfer of controls that have been applied to similar source categories or gas streams. Technologies which have not yet been applied to (or permitted for) full scale operations need not be considered as available; we do not expect the source owner to purchase or construct a process or control device that has not been demonstrated in practice. 4 Step 2 - Eliminate Technically Infeasible Options In the second step, the source-specific technical feasibility of each control option identified in step one is evaluated by answering three specific questions: a. Is the control technology available to the specific source which is undergoing the BART analysis? 3 40 CFR 51 Appendix Y 4 Federal Register 70, No. 128 (July 6, 2005):

10 The U.S. EPA states that a control technique is considered available to a specific source if it has reached the stage of licensing and commercial availability. 5 However, the U.S. EPA further states that they do not expect a source owner to conduct extended trials to learn how to apply a technology on a totally new and dissimilar source type. 6 b. Is the control technology an applicable technology for the specific source which is undergoing the BART analysis? In general, a commercially available control technology, as defined in question 1, will be presumed applicable if it has been used on the same or a similar source type. 7 If a control technology has not been demonstrated on a same or a similar source type, the technical feasibility is determined by examining the physical and chemical characteristics of the pollutant-bearing stream and comparing them to the gas stream characteristics of the source types to which the technology has been applied previously. 8 c. Are there source-specific issues/conditions that would make the control technology not technically feasible? This question addresses specific circumstances that preclude its application to a particular emission unit. This demonstration typically includes an evaluation of the characteristics of the pollutant-bearing gas stream and the capabilities of the technology 9. This also involves the identification of un-resolvable technical difficulties. However, when the technical difficulties are merely a matter of increased cost, the technology should be considered technically feasible and the technological difficulty evaluated as part of the economic analysis 10. It is also important to note that vendor guarantees can provide an indication of technical feasibility but the U.S. EPA does not consider a vendor guarantee alone to be sufficient justification that a control option will work. Conversely, the U.S. 5 Federal Register 70, No. 128 (July 6, 2005): IBID 7 IBID 8 IBID 9 IBID 10 IBID 5

11 EPA does not consider as sufficient justification that a control option or emission limit is technically infeasible. In general, the decisions on technical feasibility should be based on a combination of the evaluation of the chemical and engineering analysis and the information from vendor guarantees 11. Step 3 - Evaluate Control Effectiveness In step three, the remaining controls are ranked based on the control efficiency at the expected emission rate (post-control) as compared to the emission rate before addition of controls (pre-bart) for the pollutant of concern. Step 4 - Evaluate Impacts and Document Results In the fourth step, an engineering analysis documents the impacts of each remaining control technology option. The economic analysis compares dollar per ton of pollutant removed for each technology. In addition it includes incremental dollar per ton cost analysis to illustrate the economic effectiveness of one technology in relation to the others. Finally, Step Four includes an assessment of energy impacts and other non-air quality environmental impacts. Economic impacts were analyzed using the procedures found in the U.S. EPA Air Pollution Control Cost Manual Sixth Edition (EPA 452/B ). Equipment cost estimates from the U.S. EPA Air Pollution Control Cost Manual or U.S. EPA s Air Compliance Advisor (ACA) Air Pollution Control Technology Evaluation Model version 7.5 were used. Vendor cost estimates for this project were used when applicable. The source of the control equipment cost data are noted in each of the control cost analysis worksheets as found in Appendix A. Step 5 - Evaluate Visibility Impacts The fifth step requires a modeling analysis conducted with U.S. EPA -approved models such as CALPUFF. The modeling protocol 12, including receptor grid, meteorological data, and other factors used for this part of the analysis were provided by the Minnesota Pollution Control Agency. The model outputs, including the 98th percentile dv value and the number of days the facility contributes more than a 0.5 deciview (dv) of 11 IBID 12 MPCA. October 10, Best Available Retrofit Technology (BART) Modeling Protocol to Determine Sources Subject to BART in the State of Minnesota. 6

12 visibility impairment at each of the Class I areas, are used to establish the degree of improvement that can be reasonably attributed to each technology. The final step in the BART analysis is to select the best alternative using the results of steps 1 through 5. In addition, the U.S. EPA and MPCA guidance states that the affordability of the controls should be considered, and specifically states: 1. Even if the control technology is cost effective, there may be cases where the installation of controls would affect the viability of plant operations. 2. There may be unusual circumstances that justify taking into consideration the conditions of the plant and the economic effects requiring the use of a given control technology. These effects would include effects on product prices, the market share, and profitability of the source. Where there are such unusual circumstances that are judged to affect plant operations, you may take into consideration the conditions of the plant and the economic effects of requiring the use of a control technology. Where these effects are judged to have severe impacts on plant operations you may consider them in the selection process, but you may wish to provide an economic analysis that demonstrates, in sufficient detail for public review, the specific economic effects, parameters, and reasoning. (We recognize that this review process must preserve the confidentiality of sensitive business information). Any analysis may also consider whether competing plants in the same industry have been required to install BART controls if this information is available. 13 To complete the BART process, the analysis must establish enforceable emission limits that reflect the BART requirements and requires compliance within a reasonable period of time 14. Those limits must be developed for inclusion in the state implementation plan (SIP) that is due to U.S. EPA in December of In addition, the analysis must include requirements that the source employ techniques that ensure compliance on a continuous basis 15. which could include the incorporation of other regulatory requirements for the source, including Compliance Assurance Monitoring (40 CFR 64), Periodic Monitoring (40 CFR 70.6(a)(3)) and Sufficiency Monitoring (40 CFR 70(6)(c)(1)). If technological or economic limitations make measurement methodology for an emission unit 13 MPCA. March Guidance for Facilities Conducting a BART Analysis. Page MPCA. March Guidance for Facilities Conducting a BART Analysis. Page MPCA. March Guidance for Facilities Conducting a BART Analysis. Page 23. 7

13 infeasible, the BART limit can instead prescribe a design, equipment, work practice, operation standard, or combination of these types of standards 16. Compliance with the BART emission limits will be required within 5 years of U.S. EPA approval of the Minnesota SIP. 16 MPCA. March Guidance for Facilities Conducting a BART Analysis. Page 23. 8

14 3. Streamlined BART Analysis Within the preamble to the final federal BART rule, U.S. EPA explicitly encouraged states to include a streamlined approach for BART analyses 17. The streamlined approach will allow both states and the facilities to focus their resources on the main contributors to visibility impairment. This section of the report follows the MPCA-approved streamlined BART analysis for taconite facilities and presents the results of the streamlined approach in Table A Indurating Furnaces The indurating furnaces are sources of three visibility impairing pollutants: NO x, SO 2, and PM. Relative to NO x and SO 2, PM is not a major visibility impairing pollutant. Further, the indurating furnaces are subject to the taconite MACT standard 18 for the PM emissions. MPCA s guidance for conducting a BART review states that MPCA will rely on MACT standards to represent BART level of control for those visibility impairing pollutants addressed by the MACT standard unless there are new technologies subsequent to the MACT standard, which would lead to cost-effective increases in the level of control. [Attachment 2, March 2006, page 2]. Since the MACT standard was established very recently and becomes effective in 2006, the technology analysis is up-to-date. As a result, BART will be presumed to be equivalent to MACT for PM. A full BART analysis will be conducted for NO x and SO 2 where applicable. 3.B PM-Only Taconite MACT Emission Units In addition to the indurating furnaces, the taconite MACT standard also regulates PM emissions from Ore Crushing and Handling operations, Pellet Coolers, and Finished Pellet Handling operations. These sources operate near ambient temperature, only emit PM, and do not emit NO x or SO 2. The Ore Crushing and Handling sources and the Finished Pellet Handling sources operate with control equipment to meet the applicable MACT limits (0.008 gr/dscf for existing sources and gr/dscf for new sources). The Pellet Cooler sources are excluded from additional control under the MACT standard due to the large size of the particles and the relatively low concentration of particle emissions [FR, December 18, 2002, page 77570]. Therefore, the emissions from the pellet coolers are considered to have a negligible impact on visibility impairment, and no control requirements under the MACT standard is consistent with the intention of the BART analysis. 17 Federal Register 70, no. 128 (July 6, 2005): and CFR Subpart 63 RRRRR-NESHAPS: Taconite Iron Ore Processing 9

15 Since the MACT standard was established recently and will become effective in 2006, the technology analysis is up-to-date. Again, for these units subject to a MACT standard, BART will be presumed to be equivalent to MACT according to MPCA guidance. No further analysis will be required to establish BART for these sources. 3.C Sources of fugitive PM that are subject to MACT standards The taconite MACT standard also regulates fugitive PM emissions (fugitive PM emissions from non- MACT sources are addressed in section 3.D). These sources operate at ambient temperature, only emit PM, and do not emit NO x or SO 2. Taconite MACT fugitive sources include the following: Stockpiles (includes, but is not limited to, stockpiles of uncrushed ore, crushed ore, or finished pellets), Material Transfer Points, Plant Roadways, Tailings basins, Pellet loading areas, and Yard areas. Control of emissions from these fugitive PM sources is maintained through a fugitive control plan, as required by the MACT standard and as required by the facility s Title V air permit. The fugitive controls consist of primary controls, and contingent measures to prevent or mitigate fugitive PM emissions. The controls and measures are site specific and are appropriate to seasonal and weather conditions. Since the MACT standard was established recently and will become effective in 2006, the technology analysis is up-to-date. Again, for these units subject to a MACT standard, BART will be presumed to be equivalent to MACT according to MPCA guidance. No further analysis will be required to establish BART for these sources. 3.D Non-MACT Units and Fugitive Sources (PM only) A few sources of PM emissions and sources of fugitive PM are not subject to a MACT standard. They include units such as: Additive storage and handling Concentrate storage and handling Considering all PM emissions which are subject to the BART standard, the PM emissions from the above units typically represent less than 2.5% of PM emissions from the facility, which are subject to BART. Relative to NO x and SO 2, PM is not a major visibility impairing pollutant. 10

16 The point source emission units are typically controlled by baghouses, which achieve high levels of control for PM. Since these units already have the best control equipment for PM emissions, and since the PM emissions from these sources are small relative to the total PM emissions that are subject to the BART standard, additional control of these sources can be presumed to have minimal impact on visibility improvement in Class I areas. For the controlled sources, existing controls will be considered BART consistent with direction from MPCA in the May 18, 2006 meeting. No further analysis will be required to establish BART for these sources. The fugitive sources are controlled by measures described in the fugitive control plan as required by the facility Title V operating permit. Fugitive control plans are accepted as MACT and BACT for fugitive sources, including the recent taconite MACT standard and a BACT determination made in 2006 for the mining industry. For the fugitive sources, the existing fugitive dust control plans will be considered BART. No further analysis will be required to establish BART for these sources. 3.E Other Combustion Units United Taconite has other combustion units that are subject to BART. The combustion units are sources of three visibility impairing pollutants: NO x, SO 2, and PM. The remaining combustion sources consist of four small boilers (each less than 20 MMBtu/hr) and 10 heaters to warm work areas during cold weather. It is important to note that the emissions from the indurating furnaces represent the vast majority of emissions of all visibility impairing pollutants, with the other combustion units typically contributing less than 1% of the total emissions of each pollutant from sources that are subject to BART. The emissions from all the remaining sources are small relative to the total emissions that are subject to the BART standard. Additional control of these sources can be presumed to have minimal impact on visibility improvement in Class I areas. As directed by MPCA in the May 18, 2006 meeting, United Taconite conducted an analysis for the remaining combustion sources to demonstrate that the impact on visibility is negligible (less than or equal to 0.05 deciviews 19 ). The procedure for the analysis is detailed in section 3.F of this document. The results of that modeling demonstrated that the visibility impact of those sources is equal to 0.05 deciviews. Therefore, the sources are considered to not cause or contribute to visibility impairment in Class I areas. Therefore, the existing operations will be considered BART. No further analysis will be required to establish BART for these sources. 19 The threshold of perception of the human eye is 1-2 dv. 11

17 3.F Visibility Impact Modeling for Negligible Impacts As described in section 3.E of this document, this facility contains several sources that are assumed to have a negligible impact on visibility in Class I areas. In order to confirm this assumption, a modeling analysis was conducted to determine the impact of the emissions from these sources on visibility in Class I areas. The analysis consisted of the following: (1) Conduct air dispersion modeling for uncontrolled BART-eligible emission units and fugitive sources for the facility, as described in section 3.E above. The modeling was conducted based on the guidance in the BART Modeling Guidance as prepared by MPCA. One modeling analysis was conducted. The modeling was conducted on a focused grid (as previously agreed to with MPCA) which is based on the facility impacts as presented by MPCA in Results of Best Available Retrofit Technology (BART) Modeling to Determine Sources Subject-to-BART in the State of Minnesota (March 2006). (2) Count the days with a 98 th percentile (21 over 3-yrs, 7 each year) change in visibility greater than or equal to 0.05 deciviews (based on 10% of the facility threshold of 0.5 deciviews) at the modeled receptors with in the boundaries of each Class I area assessed over the 3-year period (3) If the modeled emission sources result in a 98 th percentile change in visibility less than or equal to 0.05 deciviews, the point and fugitive sources will be considered to not cause or contribute to visibility impairment in Class I areas. Therefore, the existing operations will be considered BART. No further analysis will be required to establish BART for these sources. (4) If the modeled emissions result in a 98 th percentile change in visibility greater than 0.05 deciviews, a full BART analysis will be conducted on the emission sources. 12

18 Table 3-1 Summary Table of BART-Eligible Units Subject to a Streamlined BART Analysis Emission Unit # Emission Unit Description Visibility- Impairing Pollutant 1 Existing Limit 1 gr/dscf Maximum Daily lbs/day 2 3.A Indurating Furnaces 40 Line 1 Pellet Induration PM SV 048 Line 2 Pellet Induration PM SV 049 Line 2 Pellet Induration PM B PM-Only Taconite MACT Emission Units 1 Crude Ore Unloading Pan Feeders PM SV008 Crude Ore Unloading PM SV009 Crude Ore Unloading PM Coarse Ore Surge Pan Feeders PM Third Stage Crusher 1 PM Third Stage Crusher 2 PM Third Stage Crusher 3 PM Third Stage Crusher 4 PM Third Stage Crusher 5 PM Third Stage Bins Conveyor PM Fourth Stage Crusher 1 PM Fourth Stage Crusher 2 PM Fourth Stage Crusher 3 PM Fourth Stage Crusher 4 PM Fourth Stage Crusher 5 PM Fourth Stage Crusher 6 PM Fourth Stage Crusher 7 PM Fourth Stage Crusher 8 PM Fourth Stage Trip/Bin/Convey PM Transfer House North PM Transfer House South PM No. 1 Rod Mill Feed PM No. 2 Rod Mill Feed PM No. 3 Rod Mill Feed PM No. 4 Rod Mill Feed PM No. 5 Rod Mill Feed PM Line 2 Grate Feed PM Line 2 Grate Discharge PM Line 2 Kiln Cooler Discharge PM Line 2 Kiln Cooler Discharge PM Line 2 Cooler Discharge Vibrating feeders PM Line 2 Pellet Cooler Exhaust PM Line 1 Grate Feed PM Line 1 Grate Discharge PM SV 116 Line 1 Kiln Cooler Discharge PM SV 045 Line 1 Kiln Cooler Discharge PM Line 1 Pellet Cooler Exhaust PM C Fugitive PM MACT Sources FS 001 Unpaved Roads - coarse tailings to tailings basin PM Fug Control Plan FS 002 Wind Erosion - Tailings Basin (Active Area) PM Fug Control Plan FS 005 Coarse Ore Surge Dump PM Fug Control Plan FS 012 Unpaved Roads - light truck traffic around plant PM Fug Control Plan FS 029 Fine Ore Surge (9F to 10 transfer) PM Fug Control Plan FS 030 Fine Ore Surge (10 to pile transfer) PM Fug Control Plan FS 031 Line 1 Pellet Transfer (21 to 21B) PM Fug Control Plan FS 033 Pellet Loadout Railcar loading PM Fug Control Plan FS 051 Line 1 Pellet transfer (21 to 22N/S) PM Fug Control Plan 3.D Non-MACT Emission Units and Fugitive Sources (PM-Only) 28 Limestone/Soda Ash Storage Bins PM Additive Unloading PM Line 2 Additive Storage Bins PM Line 2 Additive Addition PM 0.3 FS 004 Wind Erosion - Concentrate Stockpile PM Fug Control Plan 13

19 Visibility- Impairing Pollutant 1 Existing Limit 1 gr/dscf Emission Unit # Emission Unit Description Maximum Daily lbs/day 2 3.E Other Combustion Units 51 Fairlane Truck Shop Boiler PM Fairlane New Shop Boiler PM Fuel Handling Boiler #1 PM Fuel Handling Boiler #2 PM Makeup Heaters (F. Crusher Bldg) PM NA 1 PM Filterable PM only as measured by U.S. EPA Method 5 including the applicable averaging and grouping provisions, as presented in the MACT regulation, effective 10/30/ Based on baseline flow rates lbs/million Btu heat input 14

20 4. Baseline Conditions and Visibility Impacts for BART Eligible Units As indicated in U.S. EPA s final BART guidance 20, one of the factors to consider when determining BART for an individual source is the degree of visibility improvement resulting from the retrofit technology. The visibility impacts for this facility were estimated using CALPUFF, an U.S. EPA approved model recommended for comparing the visibility improvements of different retrofit control alternatives. However it is important to note that CALPUFF is a conservative model that over estimates real impacts. Therefore, although the CALPUFF baseline modeling results are important to comparing control alternatives on a relative basis they are do not accurately predict real impacts. The CALPUFF program models how a pollutant contributes to visibility impairment with consideration for the background atmospheric ammonia, ozone and meteorological data. Additionally, the interactions between the visibility impairing pollutants NO x, SO 2, PM 2.5 and PM 10 can play a large part in predicting impairment. It is therefore important to take a multi-pollutant approach when assessing visibility impacts. In order to estimate the visibility improvement resulting from the retrofit technology, the source must first be modeled at baseline conditions. Per MPCA guidance, the baseline, or pre-bart conditions, shall represent the average emission rate in units of pounds per hour (lbs/hr) and reflect the maximum 24-hour actual emissions A MPCA Subject-to-BART Modeling In order to determine which sources are Subject-to-BART in the state of Minnesota, the MPCA completed modeling of the BART-eligible emission units at various facilities in Minnesota in accordance with the Regional Haze rule. The modeling by MPCA was conducted using CALPUFF, as detailed in the Best Available Retrofit Technology (BART) Modeling Protocol to Determine Sources Subject-to-BART in the State of Minnesota, finalized in March The modeling by MPCA was conducted using emission rate information submitted by the facility. The emissions were reported in units of pounds per hour (lbs/hr) and were to reflect the maximum actual emissions during a 24-hour period under steady-state operating conditions during periods of high capacity 20 Federal Register 70, no. 128 (July 6, 2005): MPCA. March Best Available Retrofit Technology (BART) Modeling Protocol to Determine Sources Subject-to-BART in the State of Minnesota. Page 8. 15

21 utilization. The results of the modeling were presented in the document titled Results of Best Available Retrofit Technology (BART) Modeling to Determine Sources Subject-to-BART in the State of Minnesota, finalized in March The modeling conducted by MPCA demonstrated that this facility is subject-to-bart. 4.B Facility Baseline Modeling Prior to re-creating the MPCA visibility impairment model, the modeling protocol was re-revaluated. On behalf of United Taconite and the other Minnesota taconite facilities, Barr Engineering proposed changes to the modeling protocol. The changes, as submitted to MPCA on May 16, 2006 are presented in Appendix B. In addition, the maximum 24-hour emission rates were re-evaluated internally within United Taconite to verify that the emission rates represent the maximum steady-state operating conditions during periods of high capacity utilization. The maximum 24-hour emission rates were adjusted to reflect changes at the facility that occurred after the submittal of the information requested by MPCA and to reflect combustion of permitted fuels. An energy efficiency project was implemented on Line 1 to recoup waste heat in 2005 thereby reducing natural gas fuel use on Line 1. The emissions of NO x from Line 1 were adjusted to represent the new baseline conditions. Following implementation of this change, stack testing was performed and the results indicate that NO x emissions were reduced by 46%. The original baseline emissions were modeled for comparison to MPCA results. This data is summarized in Table 4-1. The full modeling analysis is presented in Appendix C. 16

22 Table 4-1 Baseline Conditions Modeling Input Data and the Basis for 24-hour Emissions Data Emissio n Unit ID 3.A Emission Unit Descripti on Line 1 Pellet Induratio n Line 2 Pellet Induratio n Line 2 Pellet Induratio n SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons Indurating Furnaces Fuel oil Solid fuel Solid Fuel NOx Max. 24- hr Actual Emission s (lb/day) 15,632 22, B PM-Only Taconite MACT Emission Units Crude Ore 1 Unloadin g Pan Feeders Crude Ore 2 Unloadin g Coarse Ore 4 Surge Pan Feeders Third Stage 5 Crusher 1 Third Stage 6 Crusher 2 Basis for NOx 24-hr Actual Emissi ons Nat gas Nat Gas Nat Gas PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) Exit Gas Tempera ture (F) , , , , , , ,800 5,244, , , ,807 5,244, , , ,794 5,244, , , ,906 5,244, , , ,024 5,244, , , ,029 5,244, , ,

23 Emissio n Unit ID Emission Unit Descripti on Third Stage Crusher 3 Third Stage Crusher 4 Third Stage Crusher 5 Third Stage Bins Conveyo r Fourth Stage Crusher 1 Fourth Stage Crusher 2 Fourth Stage Crusher 3 Fourth Stage Crusher 4 Fourth Stage Crusher 5 Fourth Stage Crusher 6 SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons NOx Max. 24- hr Actual Emission s (lb/day) Basis for NOx 24-hr Actual Emissi ons PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) Exit Gas Tempera ture (F) ,034 5,244, , , ,039 5,244, , , ,049 5,244, , , ,053 5,244, , , ,001 5,244, , , ,013 5,244, , , ,021 5,244, , , ,027 5,244, , , ,033 5,244, , , ,037 5,244, , ,

24 Emissio n Unit ID Emission Unit Descripti on Fourth Stage Crusher 7 Fourth Stage Crusher 8 Fourth Stage Trip/Bin/ Convey Transfer House North Transfer House South No. 1 Rod Mill Feed No. 2 Rod Mill Feed No. 3 Rod Mill Feed No. 4 Rod Mill Feed No. 5 Rod Mill Feed Line 2 Grate Feed Line 2 Grate Discharg e SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons NOx Max. 24- hr Actual Emission s (lb/day) Basis for NOx 24-hr Actual Emissi ons PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) Exit Gas Tempera ture (F) ,044 5,244, , , ,050 5,244, , , ,057 5,244, , , ,149 5,244, , , ,124 5,244, , , ,205 5,244, , , ,147 5,244, ,234 5,244, , , ,263 5,244, , , ,291 5,244, , , ,014 5,244, , , ,978 5,244, , ,

25 Emissio n Unit ID Emission Unit Descripti on Line 2 Kiln Cooler Discharg e Line 2 Kiln Cooler Discharg e Line 2 Cooler Discharg e Vibrating feeders Line 2 Pellet Cooler Exhaust Line 1 Grate Feed Line 1 Grate Discharg e Line 1 Kiln Cooler Discharg e Line 1 Pellet Cooler Exhaust SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons NOx Max. 24- hr Actual Emission s (lb/day) Basis for NOx 24-hr Actual Emissi ons PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) Exit Gas Tempera ture (F) ,969 5,244, , , ,933 5,244, , , ,960 5,244, , , ,958 5,244, , , ,043 5,244, , , ,244, , , ,992 5,244, , ,971 5,244, , , ,993 5,244, , ,239 1,100 20

26 Emissio n Unit ID Emission Unit Descripti on SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons 3.C Fugitive PM MACT Sources Unpaved Roads - coarse 1 tailings to tailings basin Wind Erosion - Tailings 2 Basin (Active Area) Coarse Ore 5 Surge Dump Unpaved Roads - light 12 truck traffic around plant Fine Ore Surge 29 (9F to 10 transfer) Fine Ore Surge 30 (10 to pile transfer) Line 1 Pellet 31 Transfer (21 to 21B) NOx Max. 24- hr Actual Emission s (lb/day) Basis for NOx 24-hr Actual Emissi ons PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) Exit Gas Tempera ture (F) 0.0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 77 21

27 Emissio n Unit ID Emission Unit Descripti on SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons NOx Max. 24- hr Actual Emission s (lb/day) Basis for NOx 24-hr Actual Emissi ons PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) 33 Pellet Loadout Railcar 0.0 NA NA NA NA 77 Loading 51 Line 1 Pellet 0.0 NA NA NA NA 77 Transfer 3.D Non-MACT Emission Units and Fugitive Sources (PM-Only) 28 Limeston e/soda Ash ,055 5,244, , , Storage Bins 29 Additive Unloadin ,072 5,244, , , g 30 Line 2 Additive Storage ,015 5,244, , , Bins 31 Line 2 Additive ,046 5,244, , , Addition 4 Wind Erosion - Concentr 0.0 NA NA NA NA 77 ate Stockpile 3.E Other Combustion Units Fairlane 51 Truck Shop ,039 5,244, , , Boiler 52 Fairlane New Shop ,280 5,244, , , Boiler 53 Fuel Handling Boiler # ,798 5,244, , , Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) Exit Gas Tempera ture (F) 22

28 Emissio n Unit ID Emission Unit Descripti on Fuel Handling Boiler #2 10 Makeup Heaters (F. Crusher Bldg) SO2 Max. 24-hr Actual Emissi ons (lb/day ) Basis for SO2 24-hr Actual Emissi ons NOx Max. 24- hr Actual Emission s (lb/day) Basis for NOx 24-hr Actual Emissi ons PM2.5 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM hr Actual Emissi ons PM10 Max. 24-hr Actual Emissio ns (lb/day) Basis for PM10 24-hr Actual Emissi ons Stac k No. Easting Northing Height of openin g from ground (ft) Base Elevation of Ground (ft) Stack length, width or Diamete r (ft) Flow Rate at Exit (acfm) ,802 5,244, , , Exit Gas Tempera ture (F) 23

29 4.C Facility Baseline Modeling Results The Minnesota BART modeling protocol 22 describes the CALPUFF model inputs, including the meteorological data set and background atmospheric ammonia and ozone concentrations, along with the functions of the CALPOST post processing. The CALPOST output files provide the following two methods to assess the expected post-control visibility improvement: 98 th Percentile: As defined by federal guidance and as stated in the MPCA s document which identifies the Minnesota facilities that are subject to BART 23, a source "contributes to visibility impairment if the 98 th percentile of any year s modeling results (i.e. 7 th highest day) meets or exceeds the threshold of five-tenths (0.5) of a deciview (dv) at a Federally protected Class I area receptor. Number of Days Exceeding 0.5 dv: The severity of the visibility impairment contribution, or reasonably attributed visibility impairment, can be gauged by assessing the number of days on which a source exceeds a visibility impairment threshold of 0.5 dv. A summary of the baseline visibility modeling is presented in Table 4-2. As illustrated in the table, this facility is considered to contribute to visibility impairment in Class I areas because the modeled 98 th percentile of the baseline conditions exceeds the threshold of 0.5 dv. The results of this modeling are also utilized in the post-control modeling analysis in section 6 of this document. Two baselines are shown. The first baseline represents burning natural gas on both Lines 1 and 2 and emission levels prior to installation of the heat recoup project on Line 1. The new baseline represents the emission levels after installation of the heat recoup project. The full modeling analysis is presented in Appendix C. 22 MPCA. March Best Available Retrofit Technology (BART) Modeling Protocol to Determine Sources Subject-to-BART in the State of Minnesota. Page MPCA. March Results of Best Available Retrofit Technology (BART) Modeling to Determine Sources Subject-to-BART in the State of Minnesota. 24

30 Class I Area with Greatest Impact Modeled 98 th Percentile Value (deciview) Table 4-2 Baseline Visibility Modeling Results No. of Modeled No. of Modeled days 98 th days 98 th exceeding Percentile exceeding Percentile 0.5 Value 0.5 Value deciview (deciview) deciview (deciview) No. of days exceeding 0.5 deciview Modeled 98 th Percentile Value (deciview) Combined No. of days exceeding 0.5 deciview BWCA BWCA

31 5. BART Analysis for Indurating Furnaces As presented in section 3 and Table 3-1, the only sources at United Taconite that require a full BART analysis are the indurating furnaces. The BART analysis is required for SO 2 and NO x. 5.A Indurating Furnace The primary function of taconite indurating furnaces is to convert magnetic iron concentrate to a more highly oxidized iron in the form of a pellet that is sold to metallic iron and steel production facilities. Soft or green pellets are oxidized and heat-hardened in the induration furnace. The induration process involves pellet pre-heating, drying, hardening, oxidation and cooling. The process requires large amounts of air for pellet oxidation and cooling. Process temperature control in all parts of the furnace is critical to minimize product breakage in the initial process stages, allow required oxidation reactions to occur, and adequately cool the product prior to subsequent handling steps. Directed air flow, heat recovery and fuel combustion are critical to controlling temperature and product quality in all parts of the furnace. United Taconite has grate/kiln furnaces, in which the pellets are dried on a grate and then transferred to a rotary kiln for hardening and oxidation. The pellet hardening and oxidation section of the induration furnace is designed to operate at 2,400 ºF and higher. This temperature is required to meet taconite pellet product specifications. Direct fired fuel combustion in the induration furnace is carried out at 300% to 400% excess air to provide sufficient oxygen for pellet oxidation. Air is used for combustion, pellet cooling, and as a source of oxygen for pellet oxidation. Due to the high-energy demands of the induration process, induration furnaces have been designed to recover as much heat as possible using hot exhaust gases to heat up incoming pellets. Pellet drying and preheat zones are heated with the hot gases generated in the pellet hardening/oxidation section and the pellet cooler sections. Each of these sections is designed to maximize heat recovery within process constraints. The pellet coolers are also used to preheat combustion air so more of the fuel s energy is directed to the process instead of heating ambient air to combustion temperatures. United Taconite has two grate/kiln furnaces, Line 1 and Line 2. Line 1 is permitted to burn natural gas and fuel oil. Line 2 is permitted to burn natural gas, fuel oil, coal, and petroleum coke. Both lines are controlled by wet scrubbers. Line 1 also uses caustic reagent. The wet scrubbers are designed to remove PM and would be considered a high efficiency PM wet scrubber. Since collateral SO 2 reductions occur within the existing wet scrubbers, they are considered low efficiency SO 2 26

32 scrubbers and will be evaluated as such within this BART analysis. NO x is controlled through furnace design. 5.A.i Sulfur Dioxide Controls One source of SO 2 emissions in taconite production is trace amounts of sulfur in the iron concentrate and binding agents present in the green balls. Sulfur is also present in distillate fuel oil, that is a permitted fuel for both lines, and in coal and petroleum coke, that are permitted fuels on Line 2. 5.A.i.a STEP 1 Identify All Available Retrofit Control Technologies See Appendix D for a comprehensive list of all potential retrofit control technologies that were evaluated. Many emerging technologies have been identified that are not currently commercially available. A preliminary list of technologies was submitted to MPCA on May 9 with the status of the technology as it was understood at that time. As work on this evaluation progressed, new information became apparent of the limited scope and scale of some of the technology applications. Appendix D presents the current status of the availability and applicability of each technology. 5.A.i.b STEP 2 Eliminate Technically Infeasible Options Step 1 identified the available and applicable technologies for SO 2 emission reduction. Within Step 2, the technical feasibility of the control option is discussed and determined. The following section describes retrofit SO 2 control technologies that were identified as available and applicable in the original submittal and discusses aspects of those technologies that determine whether or not the technology is technically feasible for indurating furnaces. Wet Walled Electrostatic Precipitator (WWESP) An electrostatic precipitator (ESP) applies electrical forces to separate suspended particles from the flue gas stream. The suspended particles are given an electrical charge by passing through a high voltage DC corona region in which gaseous ions flow. The charged particles are attracted to and collected on oppositely charged collector plates. Particles on the collector plates are released by rapping and fall into hoppers for collection and removal. A wet walled electrostatic precipitator (WWESP) operates on the same collection principles as a dry ESP and uses a water spray to remove particulate matter from the collection plates. For SO 2 removal, caustic is added to the water spray system, allowing the WWESP spray system to function as an SO 2 absorber. 27

33 The SO 2 control efficiency for a WWESP is dependent upon various process specific variables, such as SO 2 flue gas concentration, fuel used, and ore composition. United Taconite currently employs a wet scrubber designed for removal of particulate matter, and it performs as a low efficiency SO 2 wet scrubber. The addition of a WWESP would act as a polishing scrubber and would experience reduced control efficiency due to lower SO 2 inlet concentrations. A control efficiency as a polishing WWESP ranges from 30-80% dependent upon the process specific operating parameters. Based on the definitions contained within this report, a WWESP is considered an available technology for SO 2 reduction for this BART analysis. Wet Scrubbing (High and Low Efficiency) Wet scrubbing, when applied to remove SO 2, is generally termed flue-gas desulfurization (FGD). FGD utilizes gas absorption technology, the selective transfer of materials from a gas to a contacting liquid, to remove SO 2 in the waste gas. Crushed limestone, lime, or caustic are used as scrubbing agents. Most wet scrubbers recirculate the scrubbing solution, which minimizes the wastewater discharge flow. However, higher concentrations of solids exist within the recirculated wastewater. For a wet scrubber to be considered a high efficiency SO 2 wet scrubber, the scrubber would require designs for removal efficiency up to 95% SO 2. Typical high efficiency SO 2 wet scrubbers are packedbed spray towers using a caustic scrubbing solution. Whereas, a low efficiency SO 2 wet scrubber could be as low as 30% control efficiency. A low efficiency SO 2 scrubber could be a venturi rod scrubber design using water as a scrubbing solvent. Venturi rod scrubbers, which are frequently used for PM control at taconite facilities, will also remove some of the SO 2 from the flue gas as collateral emission reduction. Limestone scrubbing introduces limestone slurry with the water in the scrubber. The sulfur dioxide is absorbed, neutralized, and partially oxidized to calcium sulfite and calcium sulfate. The overall reactions are shown in the following equations: CaCO 3 + SO 2 CaSO 3 1/2 H2O + CO 2 CaSO 3 1/2 H 2 O + 3H 2 O + O 2 2 CaSO 4 2 H 2 O Lime scrubbing is similar to limestone scrubbing in equipment and process flow, except that lime is a more reactive reagent than limestone. The reactions for lime scrubbing are as follows: 28

34 Ca(OH) 2 +SO 2 CaSO 3 1/2 H 2 O + 1/2 H 2 O Ca(OH) 2 + SO 2 + 1/2 O 2 + H 2 O CaSO 4 2 H 2 O When caustic (sodium hydroxide solution) is the scrubbing agent, the SO 2 removal reactions are as follows: Na+ + OH- + SO 2 + Na 2 SO 3 2Na+ + 2OH- + SO 2 + Na 2 SO 3 + H 2 O Caustic scrubbing produces a liquid waste, and minimal equipment is needed as compared to lime or limestone scrubbers. If lime or limestone is used as the reagent for SO 2 removal, additional equipment will be needed for preparing the lime/limestone slurry and collecting and concentrating the resultant sludge. Calcium sulfite sludge is watery; it is typically stabilized with fly ash for land filling. The calcium sulfate sludge is stable and easy to dewater. To produce calcium sulfate, an air injection blower is needed to supply the oxygen for the second reaction to occur. The normal SO 2 control efficiency range for SO 2 scrubbers on coal-fired utility boilers with excess oxygen of 2-3% is 80% to 90% for low efficiency scrubbers and 90% to 95% for high efficiency scrubbers. The highest control efficiencies can be achieved when SO 2 concentrations are the highest. Unlike coal-fired boilers, indurating furnaces operate with maximum excess air to enable proper oxidation of the pellet. The excess air dilutes the SO 2 concentration as well as creates higher flow rates to control. Additionally, the varying sulfur concentration within the pellet causes fluctuations of the SO 2 concentrations in the exhaust gas stream. This could also impact the SO 2 control efficiency of the wet scrubber. As stated in the beginning of this section, wet scrubbers are currently in place on the furnace exhausts and are believed to remove 15 to 30% of the SO 2 in the exhaust. Taking into consideration the removal of SO 2 from the low-efficiency primary PM scrubber as well as a polishing SO 2 wet scrubber, an overall efficiency of the control train would then be 80%. United Taconite has evaluated modifying the exiting scrubber on Line 2 to determine if improvements to SO 2 removal could be accomplished. The concentration of caustic in the scrubbing water was increased to a ph of 8 for several hours and stack sampling was performed to evaluate the change in SO 2 emissions. The SO 2 emissions were not reduced. The materials of construction of the existing scrubbers would not withstand a higher ph than 7 or 8. An engineering study would be required to evaluate more extensive redesign of the existing scrubbers, such as modifying the spray, 29

35 increasing the contact time, recirculating the scrubber water, and retrofitting to allow use of a much higher ph. Quantifying the likelihood and magnitude of potential improvements to SO 2 control would not be possible without a study, so this option could not be selected as BART. Based on the information contained within this report, a polishing wet scrubber is considered an available technology for SO 2 reduction for this BART analysis. Dry Sorbent Injection (Dry Scrubbing Lime/Limestone Injection) Lime/limestone injection is a post-combustion SO 2 control technology in which pulverized lime or limestone is directly injected into the duct upstream of the fabric filter. Dry sorption of SO 2 onto the lime or limestone particle occurs and the solid particles are collected with a fabric filter. Further SO 2 removal occurs as the flue gas flows through the filter cake on the bags. The normal SO 2 control efficiency range for dry SO 2 scrubbers is 70% to 90 % for coal fired utility boilers. Induration waste gas streams are high in water content and are exhausted at or near their dew points. Gases leaving the induration furnace are currently treated for removal of particulate matter using a wet scrubber. The exhaust temperature is typically in the range of 100 F to 150 F and is saturated with water. For comparison, a utility boiler exhaust operates at 350 F or higher and is not saturated with water. Under induration furnace waste gas conditions, the baghouse filter cake would become saturated with moisture and plug both the filters and the dust removal system. Although this may be an available and applicable control option, it is not technically feasible due to the high moisture content and will not be further evaluated in this report. Spray Dryer Absorption (SDA) Spray dryer absorption (SDA) systems spray lime slurry into an absorption tower where SO 2 is absorbed by the slurry, forming CaSO 3 /CaSO 4. The liquid-to-gas ratio is such that the water evaporates before the droplets reach the bottom of the tower. The dry solids are carried out with the gas and collected with a fabric filter. When used to specifically control SO 2, the term flue-gas desulfurization (FGD) may also be used. Under induration furnace waste gas conditions, the baghouse filter cake would become saturated with moisture and plug both the filters and the dust removal system. In addition, because of the moisture in the exhaust, the lime slurry would not dry properly and it would plug up the dust collection system. Similarly to the dry sorbent injection control option, this is an available and applicable control option, but is not technically feasible due to the high moisture content. This option will not be further evaluated in this report. 30

36 Energy Efficiency Projects Energy efficiency projects provide opportunities for a company to reduce their fuel consumption, which results in lower operating costs. Typically, reduced fuel usage translates into reduced pollution emissions. United Taconite recently installed heat recuperation on Line 1 and performed stack testing to quantify the emission reductions. Testing completed in May of 2005 revealed that NO x emissions were reduced by 46%. Differences in SO 2 emissions are attributed to differences in ore composition. Therefore, this alternative will not be carried forward in this analysis for SO 2, but will be carried forward for NO x. Alternate Fuels As described within the energy efficiency description, increased price of fuel has moved companies to evaluate alternate fuel sources. These fuel sources come in all physical forms solid, liquid and gas. To achieve reduction of SO 2 emissions through alternative fuel usage, the source must use fuels with lower sulfur content. Switching fuels trades one visibility impairment pollutant (SO 2 ) for another (NO x ), as induration furnaces emit significantly less NO x when burning solid fuels. Therefore, if this option is pursued, the impact on emissions of all visibility pollutants must be quantified and the cumulative visibility impact modeled to determine the net benefit of a particular alternative fuel. It is also important to note that U.S. EPA s intent is for facilities to consider alternate fuels as an option, not to direct the fuel choice. 24 Therefore, due to the uncertainty of alternative fuel costs, the potential of replacing one visibility impairment pollutant for another, and the fact that BART is not intended to mandate a fuel switch, alternative fuels as an SO 2 air pollution control technology will not be further evaluated in this report. However, similar to energy efficiency, facilities will continue to evaluate and implement alternate fuel usage as the feasibility arises. Coal Processing Pre-combustion coal processing techniques have been proposed as one strategy to reduce uncontrolled SO 2 emissions. Coal processing technologies are being developed to remove moisture and potential contaminants from the coal prior to use. 24 Federal Register 70, no. 128 (July 6, 2005):

37 These processes typically employ both mechanical and thermal means to increase the quality of coal by removing moisture, sulfur, mercury, and heavy metals. In one process, raw coal from the mine enters a first stage separator where it is crushed and screened to remove large rock and rock material. 25 The processed coal is then passed on to an intermediate storage facility prior to being sent to the next stage in the process, the thermal process. In this stage, coal passes through pressure locks into the thermal processors where steam is injected. Moisture in the coal is released under these conditions. Mineral inclusions are also fractured under thermal stress, removing both included rock and sulfur-forming pyrites. After treatment, the coal is discharged into a second pressurized lock. The second pressurized lock is vented into a water condenser to return the processor to atmospheric pressure and to flash cool the coal. Water, removed from the process at various points, and condensed process steam are reused within the process or treated prior to being discharged. To date, the use of processed fuels has only been demonstrated with test burns in a pulverized coalfired boiler. Using processed fuels at a taconite plant would require research, test burns, and extended trials to identify potential impacts on plant systems, including the furnaces, material handling, and emission control systems. Therefore, processed fuels are not considered commercially available, and will not be analyzed further in this BART analysis. Coal crushing and drying is currently employed by United Taconite for Line 2 as an incidental option for SO 2 reduction. In the process, raw coal is crushed and screened to remove rocks and other impurities. The crushed coal is then thermally processed to remove excess moisture. Coal crushing and drying is already in use, so further reductions from this control option are not possible. Therefore, it will not be further evaluated in this report. Step 2 Conclusion Based upon the determination within Step 2, the remaining SO 2 control technologies that are available and applicable as secondary controls to the existing indurating furnace wet scrubbers are identified in Table 5-1. The technical feasibility as determined in Step 2 is also included in Table The coal processing description provided herein is based on the K-Fuel process under development by KFx, Inc. 32

38 Table 5-1 Indurating Furnace SO 2 Control Technology Availability, Applicability, and Technical Feasibility SO 2 Pollution Control Technology Available? Applicable? Technically Feasible? Wet Scrubbing (Secondary) Yes Yes Yes Modification to Existing Wet Yes Yes No Scrubber Wet Walled Electrostatic Yes Yes Yes Precipitator (WWESP) Dry sorbent injection Yes Yes No Spray Dryer Absorption Yes Yes No (SDA) Alternative Fuels Yes Yes Yes Not Required Energy Efficiency Projects Yes Yes Not Required 5.A.i.c STEP 3 Evaluate Control Effectiveness of Remaining Control Technologies Table 5-2 describes the expected control efficiency from each of the remaining feasible control options. The WWESP and high efficiency wet scrubbing control options listed in Table 5-2 would be considered polishing controls since a wet scrubber currently exists that achieves some control of SO 2. Table 5-2 Indurating Furnace SO 2 Control Technology Effectiveness SO 2 Pollution Control Technology Approximate Control Efficiency Wet Walled Electrostatic Precipitator (WWESP) 80 Secondary Wet Scrubber 60 5.A.i.d STEP 4 Evaluate Impacts and Document the Results As illustrated in Table 5-2 above, the technically feasible control remaining provide varying levels of emission reduction. Therefore, it is necessary to consider the economic, energy, and environmental impacts to better differentiate as presented below. Economic Impacts Table 5-3 details the expected costs associated with installation of the above alternatives on each stack. Equipment design was based on the maximum 24-hour emissions, vendor estimates, and U.S. 33

39 EPA cost models. Capital costs were based on a recent vendor quotation. The cost for that unit was scaled to each stack s flow rate using the 6/10 power law as shown in the following equation: Cost of equipment A = Cost of equipment B * (capacity of A/capacity of B) 0.6 Direct and indirect costs were estimated as a percentage of the fixed capital investment using U.S. EPA models and factors. Operating costs were based on 93% utilization and annual operating hours of 6600 hours for Line 1 and 8700 hours for Line 2. Operating costs of consumable materials, such as electricity, water, and chemicals were established based on the U.S. EPA control cost manual 26 and engineering experience, and were adjusted for the specific flow rates and pollutant concentrations. Due to space considerations, 60% 27 of the total capital investment was included in the costs to account for a retrofit installation. After a tour of the facility and discussions with facility staff, it was determined the space surrounding the furnaces is congested, and the area surrounding the building supports vehicle and rail traffic to transport materials to and from the building. Additionally, the structural design of the existing building would not support additional equipment on the roof. Therefore, the cost estimates provide for additional site-work and construction costs to accommodate the new equipment within the facility. A site-specific estimate for site work, foundations, and structural steel was added to arrive at the total retrofit installed cost of the control technology. The site specific estimate was based on Barr s experience with similar retrofit projects. See Appendix C for an aerial photo of the facility. The detailed cost analysis is provided in Appendix A. Table 5-3 Indurating Furnace SO 2 Control Cost Summary Control Technology Wet Walled Electrostatic Precipitator (WWESP) Line 1 Pellet Induration Line 2 Pellet Induration Installed Capital Cost (MM$) Operating Cost (MM$/yr) Annualized Pollution Control Cost ($/ton) $29,805,809 $7,441,650 $671,145 Incremental Control Cost ($/ton) $41,621,612 $17,974,280 $8,171 $22, U.S. EPA, January 2002, EPA Air Pollution Control Cost Manual, Sixth Edition. 27 U.S. EPA, CUE Cost Workbook Version 1.0. Page 2. 34

40 Control Technology Wet Scrubber (Additional) Line 1 Pellet Induration Line 2 Pellet Induration Installed Capital Cost (MM$) Operating Cost (MM$/yr) Annualized Pollution Control Cost ($/ton) $20,773,230 $3,196,687 $384,402 Incremental Control Cost ($/ton) $28,067,807 $5,545,472 $3,361 Lowest Based on the BART final rule, court cases on cost-effectiveness, guidance from other regulatory bodies, and other similar regulatory programs like Clean Air Interstate Rule (CAIR), cost-effective air pollution controls in the electric utility industry for large power plants are in the range $1,000 to $1,300 per ton removed as illustrated in Appendix E. This cost-effective threshold is also an indirect measure of affordability for the electric utility industry used by USEPA to support the BART rulemaking process. For the purpose of this taconite BART analysis, the $1,000 to $1,300 cost effectiveness threshold is used as the cutoff in proposing BART. The taconite industry is not afforded the same market stability or guaranteed cost recovery mechanisms that are afforded to the electric utility industry. Therefore, the $1,000 to $1,300 per ton removed is considered a greater business risk to the taconite industry. Thus it is reasonable to use it as a cost effective threshold for proposing BART in lieu of developing industry and site specific data. The annualized pollution control cost value was used to determine whether or not additional impacts analyses would be conducted for the technology. If the control cost was less than a screening threshold established by MPCA, then visibility modeling impacts, and energy and other impacts are evaluated. MPCA set the screening level to eliminate technologies from requiring the additional impact analyses at an annualized cost of $8,000 to $12,000 per ton of controlled pollutant 28. Therefore, all air pollution controls with annualized costs less than this screening threshold will be evaluated for visibility improvement, energy and other impacts. The incremental control cost listed in Table 5-3 is intended to represent the incremental value of each technology as compared to the technology with the next highest level of control. 28 Minnesota Pollution Control Agency (MPCA). Air Permit for Cenex Harvest States Cooperative Soybean Processing Plant, Fairmont, MN Permit

41 The cost of control for Line 1 is far above the screening threshold for carrying either technology forward in this analysis. The cost of control for Line 2 is below the screening threshold for both technologies, however the incremental cost of a WWESP is higher than the screening threshold, so a polishing wet scrubber will be carried forward in this analysis. Energy and Environmental Impacts The energy and non-air quality impacts for wet scrubbers and WWESPs are presented in Table 5-4. Table 5-4 SO 2 Control Technology Impacts Assessment Control Technology Energy Impacts Other Impacts Secondary Wet Scrubber Wet Walled Electrostatic Precipitator (WWESP) Significant pressure drop results in higher electrical requirements Additional electrical consumption due to operation of the ESP Additional water consumption and wastewater generation Increased facility waterbalance and water quality issues Ponding for scrubber discharge will be limited because of site space constraints Additional solid waste/sludge generation Process downtime/lost production during installation Additional water consumption and wastewater generation Increased facility waterbalance and water quality issues Ponding for scrubber discharge will be limited because of site space constraints Additional solid waste/sludge generation (or increased sulfate loading to wastewater for caustic reagent) Process downtime/lost production during installation Plume visibility 36

42 5.A.i.e STEP 5 Evaluate Visibility Impacts As previously stated in section 4 of this document, states are required to consider the degree of visibility improvement resulting from the retrofit technology, in combination with other factors such as economic, energy and other non-air quality impacts, when determining BART for an individual source. The baseline, or pre-bart, visibility impacts modeling was presented in section 4 of this document. This section of the report evaluates the visibility impacts of BART SO 2 control and the resulting degree of visibility improvement. Predicted 24-Hour Maximum Emission Rates Consistent with the use of the highest daily emissions for baseline, or pre-bart, visibility impacts, the post-control emissions to be used for the visibility impacts analysis should also reflect a maximum 24-hour average project emission rate. In the visibility impacts analysis for SO 2, the emissions from the sources undergoing a full BART SO 2 analysis were adjusted to reflect the projected 24-hour maximum SO 2 emission rate when applying the control technologies that met the threshold requirements of steps 1 4. The emissions from all other Subject-to-BART sources were not changed. Table 5-5 provides a summary of the modeled 24-hour maximum emission rates and their computational basis for the evaluated SO 2 control technologies. Table 5-6 provides a summary of the SO 2, NO x, and PM 10 emissions for each modeling scenario, and Table 5-7 provides a summary of the modeling input data. 37

43 Table 5-5 Post-Control Modeling Scenarios Scenario Control Technology SO 2 NOx Max 24-hour lbs/hr % Reduction Control Scenario SV # Emission Unit SO2 NOx % Reduction SV 046 EU 040 Line 1 SV 048 EU 042 Line 2 SV 049 EU 042 Line 2 Base (Wet PM Scrubbers) N G on L1 & L2; Prior to Heat Recoup on Line 1 Max 24-hour lbs/hr 0% 8 0% % 20 0% 953 0% 22 0% 1068 Base (Wet PM Scrubbers) Solid fuels on L2; Fuel Oil on L1;Heat recoup on L1 SV 046 EU 040 Line 1 0% % 651 SV 048 EU 042 Line 2 0% 346 0% 203 SV 049 EU 042 Line 2 0% 267 0% 173 Base N G on L1 & L2; with Heat (Wet PM Scrubbers) Recoup L1 SV 046 EU 040 Line 1 0% 8 46% 651 SV 048 EU 042 Line 2 0% 20 0% 953 SV 049 EU 042 Line 2 0% 22 0% Base (Wet PM Scrubbers) Solid fuels on L2; N G on L1; with Heat Recoup L1 3 4 SV 046 EU 040 Line 1 0% 8 46% 651 SV 048 EU 042 Line 2 0% 346 0% 203 SV 049 EU 042 Line 2 0% 287 0% 173 Wet Walled ESP on Solid fuels on L2; N G on L1; Line 2 Heat Recoup L1 SV 046 EU 040 Line 1 0% 8 0% 651 SV 048 EU 042 Line 2 80% 69 0% 203 SV 049 EU 042 Line 2 80% 57 0% 173 Wet SO2 Scrubber on Line 2 (secondary scrubber) Solid fuels on L2; NG on L1; Heat Recoup on L1 SV 046 EU 040 Line 1 0% SV 048 EU 042 Line 2 60% SV 049 EU 042 Line 2 60%

44 Control Scenario SV # All Emission Unit Table 5-6 Post-Control SO 2 Modeling Scenarios - Modeling Input Data Stack Easting (utm) Stack Northing (utm) Height of Opening from Ground (ft) Base Elevation of Ground (ft) Stack length, width, or Diameter (ft) Flow Rate at exit (acfm) 46 EU 040 Line EU 042 Line EU 042 Line Exit Temp ( o F) 39

45 Post-Control Visibility Impacts Modeling Results Results of the post-control visibility impacts modeling for SO 2 are presented in Table 5-8. The results summarize the 98th percentile dv value and the number of days the facility contributes more than a 0.5 dv of visibility impairment at each of the Class I areas. The comparison of the postcontrol modeling scenarios to the baseline conditions is presented in Table 5-9. It is important to note that Scenarios 6, 8, 2 and 1 are all baseline conditions using permitted fuels, as described in Table 5-5. To evaluate the impact of add-on SO 2 controls in Scenarios 3 and 4 the results are compared to Scenario 1, which uses the same fuels without SO 2 controls. As illustrated in Tables 5-8 and 5-9, the highest facility baseline visibility contribution is 4.7 dv (Scenario 6-both lines using natural gas, prior to the heat recoup project on Line 1). Installing heat recoup on Line 1 reduced that impact by 0.7 dv to 4.0 dv due to NO x emission reductions (Scenario 2-both lines using natural gas, after the heat recoup project). The facility is currently operating under Scenario 1, consisting of natural gas on Line 1 and solid fuels on Line 2. The modeled visibility impact under Scenario 1 is 2.0 dv. A polishing wet scrubber would potentially reduce the visibility contribution by 0.1 dv at a cost of over $28 million in installed capital cost and a total annual cost of over $5.5 million per year. This would result in a cost per deciview reduction of over $50 million per deciview. A WWESP on Line 2 would potentially reduce the visibility impact by 0.1 dv over $41 million in installed capital cost and a total annual cost of $16.9 million per year. This would result in a cost per deciview reduction of $117 million per deciview. Visibility impacts with NO x controls are presented in Section 6. 40

46 Scenario # Class I Area with Greatest Impact Modeled 98 th Percentile Value (deciview) 6 Base BWCA 3.5 Table 5-8 Post-Control SO 2 Modeling Scenarios - Visibility Modeling Results No. of days exceeding 0.5 deciview Modeled 98 th Percentile Value (deciview) No. of days exceeding 0.5 deciview Modeled 98 th Percentile Value (deciview) No. of days exceeding 0.5 deciview Modeled 98 th Percentile Value (deciview) Combined No. of days exceeding 0.5 deciview Base BWCA Base BWCA Base BWCA BWCA BWCA Table 5-9 Post-Control SO 2 Modeling Scenarios - Visibility Improvements Modeling Results Scenario Limiting Class I Area Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv 3 BWCA BWCA Scenarios 3 and 4 are compared to Scenario 1, consisting of solid fuels on Line 2 and after the installation of heat recoup on Line 1. 41

47 5.A.ii Nitrogen Oxide Controls To be able to control NO x it is important to understand how NO x is formed. There are three mechanisms by which NO x production occurs: thermal, fuel and prompt NO x. Fuel bound NO x is formed from fuel combustion as nitrogen compounds in the fuel are oxidized. Thermal NO x production arises from the thermal dissociation of nitrogen and oxygen molecules within the furnace. Combustion air is the primary source of nitrogen and oxygen. In taconite furnaces, thermal NO x production is a function of the residence time, free oxygen, and temperature, primarily in the flame area of the furnace. Prompt NO x is a form of thermal NO x which is generated at the flame boundary. It is the result of reactions between nitrogen and carbon radicals generated during combustion. Only minor amounts of NO x are emitted as prompt NO x. The majority of NO x is emitted as NO. Minor amounts of NO 2 are formed in the furnaces. 5.A.ii.a STEP 1 Identify All Available Retrofit Control Technologies With the understanding of how NO x is formed, available and applicable control technologies were evaluated. See Appendix D for the current status of the availability and applicability of retrofit control technologies. 5.A.ii.b STEP 2 Eliminate Technically Infeasible Options Step 1 identified the available and applicable technologies for NO x emission reduction. Within Step 2, the technical feasibility of the control option was also discussed and determined. The following describes retrofit NO x control technologies that were identified as available and applicable in the original submittal and discusses aspects of those technologies that determine whether or not the technology is technically feasible for indurating furnaces. External Flue Gas Recirculation (EFGR) External flue gas recirculation (EFGR) uses flue gas as an inert material to reduce flame temperatures thereby reducing thermal NO x formation. In an external flue gas recirculation system, flue gas is collected from the heater or stack and returned to the burner via a duct and blower. The flue gas is mixed with the combustion air and this mixture is introduced into the burner. The addition of flue gas reduces the oxygen content of the combustion air (air + flue gas) in the burner. The lower oxygen 42

48 level in the combustion zone reduces flame temperatures; which in turn reduces NO x emissions. For this technology to be effective, the combustion conditions must have the ability to be controlled at the burner tip. The normal NO x control efficiency range for EFGR is 30% to 50%. Application for EFGR technology in taconite induration is problematic for three reasons: 1. The exhaust gas in an induration furnace typically has an oxygen content that is close to ambient, or 18% oxygen, vs. a boiler which has 2% - 3% oxygen. In a boiler, the flue gas is relatively inert so it can be used as a diluent for flame temperature reduction. Taconite waste gas has much higher oxygen level; thus use of taconite waste gas for EFGR would be equivalent to adding combustion air instead of an inert gas. 2. The oxidation zone of induration furnaces needs to be above 2,400 o F in order to meet product specifications. Existing burners are designed to meet these process conditions. Application of EFGR would reduce flame temperatures. Lower flame temperatures would reduce furnace temperatures to the point that product quality could be jeopardized. 3. Application of EFGR technology increases flame length. Dilution of the combustion reactants increases the reaction time needed for fuel oxidation to occur; so, flame length increases. Therefore, application of EFGR could result in flame impingement on furnace components. That would subject those components to excessive temperatures and cause equipment failures. Although this may be an available and applicable control option, it is not technically feasible due to the high oxygen content of the flue gas and will not be further evaluated in this report. Low- NO x Burners Low- NO x burner (LNB) technology utilizes advanced burner design to reduce NO x formation through the restriction of oxygen, flame temperature, and/or residence time. LNB is typically a staged combustion process that is designed to split fuel combustion into two zones, primary combustion and secondary combustion. In the primary combustion zone of a staged fuel burner, NO x formation is limited by a rich (high fuel) condition. Oxygen levels and flame temperatures are low; this results in less NO x formation. In the secondary combustion zone, incomplete combustion products formed in the primary zone act as 43

49 reducing agents. In a reducing atmosphere, nitrogen compounds are preferentially converted to molecular nitrogen (N 2 ) over nitric oxide (NO). The estimated NO x control efficiency for LNB in high temperature applications is 10%. Low NO x burners have been installed in the preheating section of a straight grate furnace at another taconite plant. If LNB were to be applied in the indurating zone of the furnace, the reduced flame temperatures associated with LNB would adversely affect taconite pellet product quality. LNB has not been applied to the indurating zone of any grate-kiln taconite furnace. Therefore, this option is not viable. It is also important to note that there are other methods being developed for LNB which are not yet commercially available. Some incorporate various fuel dilution techniques to reduce flame temperatures; such as mixing an inert gas like CO 2 with natural gas. Water injection to cool the burner peak flame temperature is also being investigated. This technique has already been successfully used for reducing NO x emissions from gas turbines and a straight grate taconite indurating furnace in the Netherlands. The water injection technique shows promise for high temperature applications, but will not be further investigated in this report as the technology is still in the development phase and is not yet commercially available. Induced Flue Gas Recirculation Burners Induced flue gas recirculation burners, also called ultra low- NO x burners, combine the benefits of flue gas recirculation and low- NO x burner control technologies. The burner is designed to draw flue gas to dilute the fuel in order to reduce the flame temperature. These burners also utilize staged fuel combustion to further reduce flame temperature. The estimated NO x control efficiency for IFGR burners in high temperature applications is 25-50%. As noted above, taconite furnaces are designed to operate with oxygen levels near 18%. At these oxygen levels, flue gas recirculation is ineffective at NO x reduction, and it would adversely affect combustion because excessive amounts of oxygen would be injected into the flame pattern. In addition, IFGR relies on convective flow of flue gas through the burner and requires burners to be up-fired; meaning that the burner is mounted in the furnace floor and the flame rises up. Furthermore, IFGR is not feasible because the reduced flame temperatures associated with IFGR would adversely affect taconite pellet product quality. Although this may be an available and applicable control option, it is not technically feasible due to the high oxygen content of the flue gas and will not be further evaluated in this report. 44

50 Energy Efficiency Projects Energy efficiency projects provide opportunities for a company to reduce their fuel consumption, which results in lower operating costs. Typically reduced fuel usage translates into reduced pollution emissions. United Taconite recently installed heat recuperation on Line 1 and performed stack testing to quantify the emission reductions. Testing completed in May of 2005 revealed that NO x emissions were reduced 46%. This alternative will be carried forward in this analysis. Ported Kilns Ported kilns are rotary kilns that have air ports installed at specified points along the length of the kiln for process improvement. The purpose of the ports is to allow air injection into the pellet bed as it travels down the kiln bed. Ports are installed about the circumference of the kiln. Each port is equipped with a closure device that opens when it is at the bottom position to inject air in the pellet bed, and closed when it rotates out of position. The purpose of air injection is to provide additional oxygen for pellet oxidation. The oxidation reaction produces enough heat to offset the heat loss associated with air injection. Air injection reduces the overall energy use of the kiln and produces a higher quality taconite pellet. Air injection also reduces the carry over of the oxidation reaction into the pellet coolers. Ported kilns are potentially applicable to grate kilns. In the past, the technology was believed to reduce NO x formation. However, the technology vendor will not guarantee that ported kilns will reduce NO x emissions because controlling the oxygen in the firing zone is not possible due to the flow of air from the cooler 29. Any reduction in NO x would be minor and incidental to the process improvement and specific to the individual furnace. Therefore, although this may be an available and applicable technology, it does not reduce NO x emissions and will not be further evaluated in this report. Alternate Fuels As described within the energy efficiency description, increased price of fuel has moved companies to evaluate alternate fuel sources. These fuel sources come in all physical forms solid, liquid and gas. Reduction of NO x emissions through alternative fuel usage has been achieved at taconite grate kilns through the use of solid fuel. In these cases the reduction resulted due to changes from pulverized solid fuel dispersal in the kiln that resulted in lower flame temperature compared to other 29 Telephone conversation with Metso, July 18,

51 fuels. Switching from natural gas or oil to solid fuel has a potential drawback in that I can exchange one visibility impairment pollutant (NO x ) for another (SO 2 ). Therefore, if this option is pursued, the impact on emissions of all visibility pollutants must be quantified and the cumulative visibility impact modeled to determine the net benefit of a particular alternative fuel. It is also important to note that U.S. EPA s intent is for facilities to consider alternate fuels as an option, not to direct the fuel choice. 30 Process Optimization with NO x CEMS or Other Parametric Monitoring MPCA guidance lists NO x CEMS as a work practice/operational change for controlling NO x emissions 31. Parametric monitoring is a possible derivative of this alternative. Based on conversations with MPCA staff, this work practice would include process adjustments, or optimization, to minimize NO x emissions. The impact of the process adjustments would be measured using the NO x CEMS. This approach has been used in the electric utility industry to fine tune NOx emissions from boilers. One taconite plant has installed NO x CEMS to monitor emissions, but not to optimize NO x emissions through process fine tuning. That plant has experienced some reduction in NO x emissions, but these encompass multiple variables and are not attributed to process fine tuning with the NO x CEMS. Therefore, this alternative has not been demonstrated in the taconite industry. There are several concerns with utilizing process optimization as an available, applicable and technically feasible control option for the taconite industry: Taconite furnaces are designed and operated to convert magnetite to hematite in the presence of excess oxygen and require heat input to initiate the reaction which is exothermic and releases heat once initiated. Fuel combustion is only part of the process and therefore this process is different from a boiler. 30 Federal Register 70, no. 128 (July 6, 2005): MPCA. March Guidance for Facilities Conducting a BART Analysis. Page 4. 46

52 The quality of the process feed materials to the furnace is variable at some taconite operations and product quality may be compromised by attempting to fine tune heat input to minimize NOx formation. At some operations, the operating parameters which generally influence the rate of NO x generation such as flame temperature, fuel usage and excess air are relatively constant during operation of the furnace, independent of process operation variability. This indicates that NO x formation may not be dependent upon controllable operating parameters. In the absence of controllable parameters, process optimization would not be effective at controlling NO x emissions. Based upon this information, there is no indication that further emission reductions would be achieved through process optimization, using NO x CEMS or other parametric monitoring, as a control technology. Therefore, process optimization as a control option will not be evaluated further in this report. Post Combustion Controls NO x can be controlled using add-on systems located downstream of the furnace area of the combustion process. The two main techniques in commercial service include the selective non catalytic reduction (SNCR) process and the selective catalytic reduction (SCR) process. There are a number of different process systems in each of these categories of control techniques. In addition to these treatment systems, there are a large number of other processes being developed and tested on the market. These approaches involve innovative techniques of chemically reducing, absorbing, or adsorbing NO x downstream of the combustion chamber. Examples of these alternatives are nonselective catalytic reduction (NSCR) and Low Temperature Oxidation (LTO). Each of these alternatives is described below. Non-Selective Catalytic Reduction (NSCR) A non-selective catalytic reduction (NSCR) system is a post combustion add-on exhaust gas treatment system. NSCR catalyst is very sensitive to poisoning; so; NSCR is usually applied primarily in natural gas combustion applications. NSCR is often referred to as three-way conversion catalyst because it simultaneously reduces NO x, unburdened hydrocarbons (UBH), and carbon monoxide (CO). Typically, NSCR can achieve NO x emission reductions of 90 percent. In order to operate properly, the combustion process must be near 47

53 stoichiometric conditions. Under this condition, in the presence of a catalyst, NO x is reduced by CO, resulting in nitrogen (N 2 ) and carbon dioxide (CO 2 ). The most important reactions for NO x removal are: 2CO + 2NO 2CO 2 + N 2 (1) [UBH] + NO N 2 + CO 2 + H 2 O (2) NSCR catalyst has been applied primarily in clean combustion applications. This is due in large part to the catalyst being very sensitive to poisoning, making it infeasible to apply this technology to the indurating furnace. In addition, we are not aware of any NSCR installations on taconite induration furnaces or similar combustion equipment. Therefore, this technology will not be further evaluated in this report. Selective Catalytic Reduction and Regenerative Selective Catalytic Reduction SCR is a post-combustion NO x control technology in which ammonia (NH 3 ) is injected into the flue gas stream in the presence of a catalyst. NO x is removed through the following chemical reaction: 4 NO + 4 NH 3 + O 2 4 N H 2 O (1) 2 NO NH 3 + O 2 3 N H 2 O (2) A catalyst bed containing metals in the platinum family is used to lower the activation energy required for NO x decomposition. SCR requires a temperature range of about 570 F 850 F for a normal catalyst. At temperature exceeding approximately 670ºF, the oxidation of ammonia begins to become significant. At low temperatures, the formation of ammonium bisulfate causes scaling and corrosion problems. A high temperature zeolite catalyst is also available; it can operate in the 600 F 1000 F temperature range. However, these catalysts are very expensive. Ammonia slip from the SCR system is usually less than 3 to 5 ppm. The emission of ammonia increases during load changes due to the instability of the temperature in the catalyst bed as well as at low loads because of the low gas temperature. Regenerative Selective Catalytic Reduction (RSCR) applies the Selective Catalytic Reduction (SCR) control process as described below with a preheat process step to reheat the flue gas stream up to SCR catalyst operating temperatures. The preheating process combines use of a thermal heat sink 48

54 (packed bed) and a duct burner. The thermal sink recovers heat from the hot gas leaving the RSCR and then transfers that heat to gas entering the RSCR. The duct burner is used to complete the preheating process. RSCR operates with several packed bed/scr reactor vessels. Gas flow alternates between vessels. Each of the vessels alternates between preheating/treating and heat recovery. The benefits of RSCR are: Its high energy efficiency allows it to be used after SO 2 and particulate controls. RSCR has a thermal efficiency of 90% - 95% vs. standard heat exchangers which have a thermal efficiency of 60% to 70%. Application of RSCR after SO 2 and PM controls significantly reduces the potential for problems associated with plugging and catalyst poisoning and deactivation. There are several concerns about the technical feasibility and applicability of RSCR on an indurating furnace: The composition of the indurating furnace flue gas is significantly different from the composition of the flue gas from the boilers that utilize RSCR; The taconite dust is highly erosive and can cause significantly equipment damage. RSCR has a number of valves which must be opened and closed frequently to switch catalyst/heat recovery beds. These valves could be subject to excessive wear in a taconite application due to the erosive nature of the taconite dust; RSCR has not been applied downstream of a wet scrubber. Treating a stream saturated with water may present design problems in equipment sizing for proper heat transfer and in corrosion protection; RSCR catalyst had been shown to oxidize mercury. Oxidized mercury can be absorbed by the local environment and have adverse impact. The impact of RSCR on mercury emissions needs to be studied to determine whether or not mercury oxidation is a problem and to identify mitigation methods if needed. 49

55 To date, RSCR has been applied to wood-fired utility boilers. Application of this technology has not been applied to taconite induration furnaces, to airflows of the magnitude of taconite furnace exhausts, nor to exhaust streams with similar, high moisture content. Using RSCR at a taconite plant would require research, test runs, and extended trials to identify potential issues related to catalyst selection, and impacts on plant systems, including the furnaces and emission control systems. It is not reasonable to assume that vendor guarantees of performance would be forthcoming in advance of a demonstration project. The timeline required to perform such a demonstration project would likely be two years to develop and agree on the test plan, obtain permits for the trial, commission the equipment for the test runs, perform the test runs for a reasonable study period, and evaluate and report on the results. The results would not be available within the time window for establishing emission limits to be incorporated in the state implementation plan (SIP) by December Recalling U.S. EPA s intention regarding available technologies to be considered for BART, as mentioned in Section 2.B, facility owners are not expected to undergo extended trials in order to learn how to apply a control technology to a completely new and significantly different source type. Therefore, RSCR is not considered to be technically feasible, and will not be analyzed further in this BART analysis. SCR with reheat through a conventional duct burner (rather than using a regenerative heater) has been successfully implemented more widely and in higher airflow applications and will be carried forward in this analysis as available and applicable technology that is reasonably expected to be technically feasible. Low Temperature Oxidation (LTO) The LTO system utilizes an oxidizing agent such as ozone to oxidize various pollutants including NO x. In the system, the NO x in the flue gas is oxidized to form nitrogen pentoxide (equations 1, 2, and 3). The nitrogen pentoxide forms nitric acid vapor as it contacts the water vapor in the flue gas (4). Then the nitric acid vapor is absorbed as dilute nitric acid and is neutralized by the sodium hydroxide or lime in the scrubbing solution forming sodium nitrate (5) or calcium nitrate. The nitrates are removed from the scrubbing system and discharged to an appropriate water treatment system. Commercially available LTO systems include Tri-NO x and LoTOx. NO + O 3 NO 2 + O 2 (1) NO 2 + O 3 NO 3 + O 2 (2) NO 3 + NO 2 N 2 O 5 (3) 50

56 N 2 O 5 + H 2 O 2HNO 3 (4) HNO 3 + NaOH NaNO 3 + H 2 O (5) Low Temperature Oxidation (Tri-NO x ) This technology uses an oxidizing agent such as ozone or sodium chlorite to oxidize NO to NO 2 in a primary scrubbing stage. Then NO 2 is removed through caustic scrubbing in a secondary stage. The reactions are as follows: O 3 + NO O 2 + NO 2 (1) 2NaOH + 2NO 2 + ½ O 2 2NaNO 3 + H2O (2) Tri-NO x is a multi-staged wet scrubbing process in industrial use. Several process columns, each assigned a separate processing stage, are involved. In the first stage, the incoming material is quenched to reduce its temperature. The second, oxidizing stage, converts NO to NO 2. Subsequent stages reduce NO 2 to nitrogen gas, while the oxygen becomes part of a soluble salt. Tri-NO x is typically applied at small to medium sized sources with high NO x concentration in the exhaust gas (1,000 ppm NO x ). NO x concentrations in the exhaust at United Taconite are typically less than 200 ppm. Therefore, Tri-NO x is not applicable to taconite processing and will not be analyzed further in this BART analysis. Low Temperature Oxidation (LoTOx ) BOC Gases Lo-TOx is an example of a version of an LTO system. LoTOx technology uses ozone to oxidize NO to NO 2 and NO 2 to N 2 O 5 in a wet scrubber (absorber). This can be done in the same scrubber used for particulate or sulfur dioxide removal, The N 2 O 5 is converted to HNO 3 in a scrubber, and is removed with lime or caustic. Ozone for LoTOx is generated on site with an electrically powered ozone generator. The ozone generation rate is controlled to match the amount needed for NO x control. Ozone is generated from pure oxygen. In order for LoTOx to be economically feasible, a source of low cost oxygen must be available from a pipeline or on site generation. The first component of the technical feasibility review includes determining if the technology would apply to the process being reviewed. This would include a review and comparison of the chemical and physical properties required. Although it appears that the chemistry involved in the LTO 51

57 technology may apply to an indurating furnace, the furnace exhaust contains other ore components that may participate in side reactions. This technology has not been demonstrated on a taconite pellet indurating furnace. This raises uncertainties about how or whether the technology will transfer to a different type of process. The second component of the technical feasibility review includes determining if the technology is commercially available. Evaluations of LTO found that it has only been applied to small to medium sized coal or gas fired boiler applications, and has never been demonstrated on a large-scale facility. For example, the current installations of LoTOx are on sources with flue gas flow rates from ,000 acfm, which is quite small, compared to the indurating furnace flue gas flow rates of up to 726,000 acfm. Therefore, the application of LTO would be more than an order of magnitude larger than the biggest current installation. This large scale-up is contrary to good engineering practices and could be problematic in maintaining the current removal efficiencies. In addition, only two of BOC s LoTOx installations are fully installed and operational applications. Therefore, although this is an emerging technology, the limited application means that it has not been demonstrated to be an effective technology in widespread application. There are several other concerns about the technical feasibility and applicability of LTO on an indurating furnace: The composition of the indurating furnace flue gas is significantly different from the composition of the flue gas from the boilers and process heaters that utilize LTO; The taconite dust in the flue gas is primarily magnetite (Fe 3 O 4 ) which would react with the ozone to form hematite (Fe 2 O 3 ); since the ozone injection point would be before the scrubber, there can be more than 400 pounds per hour of taconite dust in the flue gas which could consume a significant amount of the ozone being generated which may change the reaction kinetics; consequently, this would necessitate either an increase in the amount of ozone generated or a decrease in the estimated control efficiency; The ozone that would be injected into the flue gas would react with the SO 2, converting the material to SO3 which could result in the generation of sulfuric acid mist from the scrubber; 52

58 Since LTO has not been installed at a taconite plant, it is likely that the application of LTO to an indurating furnace waste gas could present technical problems which were not encountered, or even considered, in the existing LTO applications; An LTO system at a taconite facility would also be a source of nitrate discharge to the tailings basin which would change the facility water chemistry which could cause operational problems and would likely cause additional problems with National Pollutant Discharge Elimination System (NPDES) discharge limits and requirements. Application of this technology has not been applied to taconite induration furnaces, to airflows of the magnitude of taconite furnace exhausts, nor to exhaust streams with similar, high moisture content. Using LTO at a taconite plant would require research, test runs, and extended trials to identify potential issues related to design for high airflows and impacts on plant systems, including the furnaces and emission control systems. It is not reasonable to assume that vendor guarantees of performance would be forthcoming in advance of a demonstration project. The timeline required to perform such a demonstration project would likely be two years to develop and agree on the test plan, obtain permits for the trial, commission the equipment for the test runs, perform the test runs for a reasonable study period, and evaluate and report on the results. The results would not be available within the time window for establishing emission limits to be incorporated in the state implementation plan (SIP) by December Recalling U.S. EPA s intention regarding available technologies to be considered for BART, as mentioned in Section 2.B, facility owners are not expected to undergo extended trials in order to learn how to apply a control technology to a completely new and significantly different source type. Consequently, the technical feasibility of LTO on an indurating furnace is technically infeasible for this application and will not be evaluated further. Step 2 Conclusion Based upon the determination within Step 2, the remaining NO x control technologies that are available and applicable to the indurating furnace process are identified in Table The technical feasibility as determined in Step 2 is also included in Table

59 Table 5-10 Indurating Furnace NO x Control Technology Availability, Applicability, and Technical Feasibility NO x Pollution Control Technology Available? Applicable? External Flue Gas Recirculation (EFGR) Technically Feasible? Yes No No Low- NO x Burners Yes Yes No Induced Flue Gas Recirculation Burners Yes Yes No Energy Efficiency Projects Yes Yes Yes Ported Kilns (Applies to Grate Kilns Only) Yes Yes No Alternative Fuels Yes Yes Process Optimization using NO x CEMS Non-Selective Catalytic Reduction (NSCR) Selective Catalytic Reduction (SCR) with conventional reheat Yes, Not Required Yes No No Yes No No Yes Yes Yes Regenerative SCR Yes No No Selective Non-Catalytic Reduction (SNCR) Low Temperature Oxidation (LTO) Yes No No Yes No No 5.A.ii.c STEP 3 Evaluate Control Effectiveness of Remaining Control Technologies Table 5-11 describes the expected control efficiency from each of the remaining technically feasible control options as identified in Step 2. Table 5-11 Indurating Furnace NO x Control Technology Effectiveness NO x Pollution Control Technology SCR with Conventional Reheat Approximate Control Efficiency 80% 5.A.ii.d STEP 4 Evaluate Impacts and Document the Results Table 5-12 summarizes the expected costs associated with installation of low NO x burners and SCR with conventional reheat. Capital costs were calculated based on the maximum 24-hour emissions, 54

60 U.S. EPA cost models, and vendor estimates. Vendor estimates for capital costs based on a specific flow rate were scaled to each stack s flow rate using the 6/10 power law to account for the economy of scale. Operating costs were based on 93% utilization and annual operating hours of 6600 hours for Line 1 and 8700 hours for Line 2. Operating costs were proportionally adjusted to reflect site specific flow rates and pollutant concentrations. After a tour of the facility and discussions with facility staff, it was determined the space surrounding the furnaces is congested and the area surrounding the building supports vehicle and rail traffic to transport materials to and from the building. A site-specific estimate for site-work, foundations, and structural steel was added based upon the facility site to arrive at the total retrofit installed cost of the control technology. The site specific estimate was based on Barr s experience with recent actual retrofit costs. See Appendix C for a site plan of the facility. Additionally, the structural design of the existing building would not support additional equipment on the roof. The detailed cost analysis is provided in Appendix A. Table 5-12 NO x Control Cost Summary Installed Capital Cost (MM$) Operating Cost (MM$/yr) Annualized Pollution Control Cost ($/ton) Incremental Control Cost ($/ton) Control Technology Selective Catalytic Reduction (SCR) with Reheat Line 1 Pellet Induration $44,591,384 $23,923,236 $13,659 NA Line 2 Pellet Induration solid fuels Line 2 Pellet Induration natural gas $70,762,832 $35,189,143 $22,017 NA $72,222,298 $37,770,123 $4,736 NA Based on the BART final rule, court cases on cost-effectiveness, guidance from other regulatory bodies, and other similar regulatory programs like Clean Air Interstate Rule (CAIR), cost-effective air pollution controls in the electric utility industry for large power plants are in the range $1,000 to $1,300 per ton removed as illustrated in Appendix E. This cost-effective threshold is also an indirect measure of affordability for the electric utility industry used by USEPA to support the BART rulemaking process. For the purpose of this taconite BART analysis, the $1,000 to $1,300 cost effectiveness threshold is used as the cutoff in proposing BART. The taconite industry is not afforded the same market stability or guaranteed cost recovery mechanisms that are afforded to the electric utility industry. Therefore, the $1,000 to $1,300 per ton removed is considered a greater 55

61 business risk to the taconite industry. Thus it is reasonable to use it as a cost effective threshold for proposing BART in lieu of developing industry and site specific data. The annualized pollution control cost value was used to determine whether or not additional impacts analyses would be conducted for the technology. If the control cost was less than a screening threshold established by MPCA, then visibility modeling impacts, and energy and other impacts are evaluated. MPCA set the screening level to eliminate technologies from requiring the additional impact analyses at an annualized cost of $8,000 to $12,000 per ton of controlled pollutant 32. Therefore, all air pollution controls with annualized costs less than this screening threshold will be evaluated for visibility improvement, energy and other impacts. The incremental control cost column in Table 5-6 is intended to present the incremental value of each technology as compared to the technology with the next most effective alternative. Since none of the NO x reduction technologies are cost effective, the incremental cost is not applicable. The cost of NO x control using SCR for Line 1 and Line 2 using solid fuels are over $12,000 per ton removed, which is above the screening threshold for further consideration. When Line 2 uses natural gas, the cost per ton removed is $4736, which is below the $12,000 screening level, but is more than a factor of three times the cost effectiveness threshold of $1000-$1300. this alternative is carried forward to analyze the energy and environmental impacts as well as visibility impacts. Energy and Environmental Impact The energy and non-air quality impacts for the remaining alternatives are presented in Table Table 5-13 NO x Control Technology Impacts Assessment Control Technology Energy Impacts Other Impacts SCR with Reheat Reheat would require additional natural gas use. Ammonia slip, which contributes to regional haze. Ammonia reacts with NO x to form ammonium nitrate and SO 2 to form ammonium sulfate particles. Ammonium sulfate is hygroscopic and bonds with water in the air to grow large particles. The 32 Minnesota Pollution Control Agency (MPCA). Air Permit for Cenex Harvest States Cooperative Soybean Processing Plant, Fairmont, MN Permit

62 large particles formed by ammonium sulfate disproportionately contribute to visibility impairment. Ammonia emissions will increase condensable PM emissions that will have possible PSD permitting implications. Ammonia deposition onto nearby lakes and waters of the state and contribute nutrients and undesirable biological growth. Additional safety and regulatory concerns associated with ammonia storage on site. Possible oxidation of elemental mercury. 5.A.ii.e STEP 5 Evaluate Visibility Impacts As previously stated in section 4 of this document, states are required to consider the degree of visibility improvement resulting from the retrofit technology, in combination with other factors such as economic, energy and other non-air quality impacts, when determining BART for an individual source. The baseline, or pre-bart, visibility impacts modeling was presented in section 4 of this document. This section of the report evaluates the visibility impacts of BART NO X control and the resulting degree of visibility improvement. Predicted 24-Hour Maximum Emission Rates Consistent with the use of the highest daily emissions for baseline, or pre-bart, visibility impacts, the post-control emissions to be used for the visibility impacts analysis should also reflect a maximum 24-hour average project emission rate. In the visibility impacts NO x modeling analysis, the emissions from the sources undergoing a full BART NO x analysis were adjusted to reflect the projected 24-hour maximum NO x emission rate when applying the control technologies that met the threshold requirements of steps 1 4. The emissions from all other Subject-to-BART sources were not changed. Table 5-14 provides a summary of the modeled 24-hour maximum emission rates and their computational basis for the evaluated NO x control technologies. Table 5-15 provides a summary of the SO 2, NO x, and PM 10 emissions for each modeling scenario, and Table 5-16 provides a summary of the modeling input data. 57

63 Table 5-14 Post-Control NO x Modeling Scenarios Control Scenario SV # Emission Unit Descripition NO x Control Technology Base: NG on L1 & L2; Prior to Heat Recoup on Line 1 NO x % Reduction 24-hr Max (lb/hr) SV 046 Line 1 SV 046 0% 1208 SV 048 Line 2 SV 048 0% 953 SV 049 Line 2 SV 049 0% 1068 Base: NG on L1 & L2; with Heat Recoup on Line 1 SV 046 Line 1 SV % 651 SV 048 Line 2 SV 048 0% 953 SV 049 Line 2 SV 049 0% 1068 NG on L1 & L2; with Heat Recoup on Line 1; SCR on Line 2 SV 046 Line 1 SV % 651 SV 048 Line 2 SV % 191 SV 049 Line 2 SV % 214 Table 5-15 Post-Control NO X Modeling Scenarios Emissions Data Control Scenario SV # Emission Unit Scenario Control Technology SO2 NOx Max Max % % SO2 NOx 24-hour 24-hour Reduction Reduction lbs/hr lbs/hr Prior to Heat Recoup on Line 1 Old Base: Prior to Heat Recoup on Line 1 SV 046 Line 1 SV 046 0% 8 0% 1208 SV 048 Line 2 SV 048 0% 20 0% 953 SV 049 Line 2 SV 049 0% 22 0% 1068 Base (Wet PM Scrubbers) New Base with Heat Recoup on Line 1 SV 046 Line 1 SV % 8 46% 651 SV 048 Line 2 SV % 20 0% 953 SV 049 Line 2 SV % 22 0% 1068 Base (Wet PM Scrubbers) NG on L1&L2; with Heat Recoup on Line 1; SCR on Line 2 SV 046 Line 1 SV % 8 46% 651 SV 048 Line 2 SV % 20 80% 191 SV 049 Line 2 SV % 22 80%

64 Table 5-16 Post-Control NO X Modeling Scenarios - Visibility Modeling Input Data Control Scenario SV # Emission Unit Descripition Stack Easting (utm) Stack Northing (utm) Height of Opening from Ground (ft) Base Elevation from Ground (ft) Stack length, width or Diameter (ft) Flow Rate at Exit (acfm) Exit Temp ( o F) All 46 Line 1 SV Line 2 SV Line 2 SV Post-Control Visibility Impacts Modeling Results Results of the post-control visibility impacts modeling for NO X are presented in Table The results summarize 98th percentile dv value and the number of days the facility contributes more than a 0.5 dv of visibility impairment at each of the Class I areas. The comparison of the post-control modeling scenarios to the baseline conditions is presented in Table As illustrated in tables 5-17 and 5-18, the facility baseline visibility contribution is 4.7 dv. Installing heat recoup on Line 1 reduced that impact by 0.7 dv to 4.0 dv. Installing SCR on Line 2 when natural gas is burned could potentially reduce the visibility contribution by 2.1 dv compared to Scenario 2, which includes the heat recoup project already installed on Line 1. Line 2 currently uses solid fuels (Scenario 1, see Table 5-5 through 5-9). The visibility improvement reflected in Scenario 7 is realized when Line 2 uses natural gas, which is a permitted scenario, but does not reflect current operations. Visibility impacts with SO 2 controls are presented in section 6. 59

65 Table 5-17 Post-Control NO X Modeling Scenarios - Visibility Modeling Results Scenario Limiting Class I Area Modeled 98%ile dv Modeling Results # days over 0.5 dv Modeled 98%ile dv # days over 0.5 dv Modeled 98%ile dv # days over 0.5 dv Modeled 98%ile dv # days over 0.5 dv 6 BWCA BWCA BWCA Table 5-18 Post-Control NO X Modeling Scenarios - Visibility Improvements Scenario Limiting Class I Area Improved Modeled 98 th Percentile Value ( -dv) Modeling Results Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv 2 BWCA BWCA

66 6. Visibility Impacts As previously stated in section 4 of this document, states are required to consider the degree of visibility improvement resulting from the retrofit technology, in combination with other factors such as economic, energy and other non-air quality, when determining BART for an individual source. The baseline, or pre-bart, visibility impacts modeling was presented in section 4 of this document. The visibility impacts of individual control technologies were presented in Step 5 of section 5 of this document. This section of the report evaluates the various BART control scenarios utilizing both SO 2 and NO x controls, and estimates the resulting degree of visibility improvement. 6.A Post-Control Modeling Scenarios Steps 1-4 of the BART analysis identified the control technologies that were: Available and applicable; Technically feasible; and Below the screening cost threshold for further BART analysis. Step 5 of the BART analysis evaluated the visibility impacts of each of the control technologies that met the requirements of the screening analysis of steps 1-4. The interactions between the visibility impairing pollutants NO x, SO 2 and PM 10 can play a large part in predicting impairment. It is therefore important to take a multi-pollutant approach when assessing visibility impacts. Accordingly, this visibility improvement analysis evaluates several operating control scenarios that account for the various combinations of available NO x controls. In addition, two site-specific scenarios were developed so that the evaluation includes other operating scenarios and conditions that would improve visibility impairment. The post-control modeling scenarios, including those presented in Step 5 of section 5, are presented in Table 6-1. The modeling scenario stack parameters are presented in Table B Post-Control Modeling Results Results of the post-control modeling scenarios are presented in Table 6-3. The results summarize 98th percentile dv value and the number of days the facility contributes more than a 0.5 dv of visibility impairment at each of the Class I areas. The comparison of the post-control modeling 61

67 scenarios to the baseline conditions is presented in Table 6-4. Additionally, Table 6-5 illustrates the scenarios on a $/dv basis. As mentioned in section 5.B.ii.e, the energy efficiency projects implemented at United Taconite will reduce its contribution to visibility impacts at Class I areas from 4.7 dv to 4.0 dv, which is a reduction of 15%. These reductions have already been implemented. They represent a significant contribution toward MPCA s default glide-path goal of reducing impacts by 1.67% per year. 62

68 Table 6-1 Post-Control Modeling Scenarios Control Scenario 6 SV # Emission Unit SO2 Base (Wet PM Scrubbers) Scenario Control Technology SO2 NOx Particulate Matter Max Max PM10 PM2.5 % % NOx 24-hour 24-hour Max 24-hr Max 24-hr Reduction Reduction lbs/hr lbs/hr lbs/hr lbs/hr Base:NG on L1 & L2;Prior to Heat Recoup on L1 SV 046 Line 1 SV 046 0% 8 0% 1208 SV 048 Line 2 SV 048 0% 20 0% 953 SV 049 Line 2 SV 049 0% 22 0% PMcoarse Max 24-hr lbs/hr Base (Wet PM Scrubbers) Base: Solid fuel on L2; fuel oil on L1; heat recoup on L1 SV 046 Line 1 SV % % 651 SV 048 Line 2 SV % 346 0% 203 SV 049 Line 2 SV % 287 0% 173 Base (Wet PM Scrubbers) Base:NG on L1&L2; heat recoup on L1 SV 046 Line 1 SV % 8 46% 651 SV 048 Line 2 SV % 20 0% 953 SV 049 Line 2 SV % 22 0% 1068 Base (Wet PM Scrubbers) Base: Solid fuel on L2; NG on L1; heat recoup on L1 SV 046 Line 1 SV % 8 46% 651 SV 048 Line 2 SV % 346 0% 203 SV 049 Line 2 SV % 287 0% 173 Wet Walled ESP on Line 2 Base Solid Fuel on L2; NG on L1; heat recoup on L1 SV 046 Line 1 SV 046 0% 8 0%

69 Control Scenario 4 Scenario Control Technology SO2 NOx Particulate Matter Max Max PM10 PM2.5 PMcoarse SV # Emission Unit % % SO2 NOx 24-hour 24-hour Max 24-hr Max 24-hr Max 24-hr Reduction Reduction lbs/hr lbs/hr lbs/hr lbs/hr lbs/hr 81 SV 048 Line 2 SV % 69 0% 203 SV 049 Line 2 SV % 57 0% 173 Wet SO2 Scrubber on Line 2 (secondary scrubber) Base Solid Fuel on L2; NG on L1; heat recoup on L1 SV 046 Line 1 SV 046 0% SV 048 Line 2 SV % SV 049 Line 2 SV % Base (Wet PM Scrubbers) Base: NG on L1&L2; heat recoup on L1; SCR on L2 SV 046 Line 1 SV % 8 46% 651 SV 048 Line 2 SV % 20 80% 191 SV 049 Line 2 SV % 22 80% Table 6-2 Post-Control NO X Modeling Scenarios - Modeling Input Data Control Scenario SV # All Emission Unit Stack Easting (utm) Stack Northing (utm) Height of Opening from Ground (ft) Base Elevation of Ground (ft) Stack length, width, or Diameter (ft) Flow Rate at exit (acfm) 46 Line 1 SV Line 2 SV Line 2 SV Exit Temp ( o F) 64

70 Table 6-3 Post-Control Modeling Scenarios - Visibility Modeling Results Scenario # 6 Base 8 Base Class I Area with Greatest Impact Modeled 98 th Percentil e Value (deciview ) No. of days exceedin g 0.5 deciview Modeled 98 th Percentile Value (deciview) No. of days exceeding 0.5 deciview Modeled 98 th Percentile Value (deciview) No. of days exceeding 0.5 deciview Modeled 98 th Percentile Value (deciview) Combined No. of days exceeding 0.5 deciview BWCA BWCA Base BWCA Base BWCA WWESP BWCA Secondary Wet Scrubber BWCA SCR with Nat Gas on Line 2 BWCA

71 Table 6-4 Post-Control Modeling Scenarios - Visibility Modeling Results Modeling Results Limiting Class I Area Scenario 2 BWCA Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv Improved Modeled 98 th Percentile Value ( -dv) Decreased No. of Days exceeding 0.5 dv BWCA BWCA BWCA

72 Table 6-5 Post-Control Modeling Results Dollars per Deciview Improvement Scenario SO2 NOx Base (Wet PM Scrubbers) Wet Walled ESP on Line 2 Wet SO 2 Scrubber on Line 2 (secondary scrubber) Base (Wet PM Scrubbers) Annualized Cost ($/yr) Improved Modeled 98th Percentile Value ( -dv) $/dv Solid fuels on L2; NG on L1; Heat Recoup on L Solid fuels on L2; NG on L1; Heat Recoup on L1 $16,872, $117,990,010 Solid fuels on L2; NG on L1; Heat Recoup on L1 $5,590, $52,247,645 Nat Gas with SCR on L2; NG with heat recoup on L1 $37,770, $17,985,773 C:\Documents and Settings\bth\My Documents\@Projects\BART\Utac\REV Utac BART Report_DRAFT ( ).doc 67

73 7. Select BART BART for United Taconite is determined to be as described below. For SO 2, polishing add-on controls are not cost effective. Therefore, BART is determined to be existing controls. The corresponding SO 2 emissions limit is 2 lb per million Btu heat input for Line 1 when liquid fuel is burned. For Line 2, the BART limit is 4 lb per million Btu heat input when solid fuel is burned or 2 lb per million Btu heat input when liquid fuel is burned. For NO x, the reduction due to the heat recuperation project on Line 1is selected as BART. Other addon controls for NO x are not cost effective. The NO x reductions for the heat recuperation project have been established through stack testing on Line 1. This represents a reduction in NO x emissions by 557 lb per hr. Heat recoup is integral to the process, therefore no limitation is required. For PM, requirements compelled by the October 30, 2006 MACT standard constitute BART. The corresponding emissions limits are equivalent to the limits identified in Table 3-1. The overall improvement in visibility contribution is a reduction by 15%. This represents a significant contribution to the default glide path of reducing impacts by 1-2% per year. The schedule for implementation of controls is by October 30, 2006, well in advance of the 5-year time-frame required for BART implementation. 68

74 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 1 >: Cost Summary NO x Control Cost Summary Control Technology Control Eff % Controlled Emissions T/y Emission Reduction T/yr Installed Capital Cost $ Total Annual Cost $/yr Pollution Control Cost $/ton Incremental Control Cost $/ton Selective Catalytic Reduction (SCR) + Reheat Line 1 Pellet Induration 80% $44,591,384 $23,923,236 $13,659 NA Line 2 Pellet Induration 80% $70,762,832 $35,189,143 $22,017 NA Line 2 Pellet Induration NG 80% $72,222,298 $37,770,123 $4,736 NA Energy Efficiency Project - Heat Recoup on Line 1 Line 1 Pellet Induration 46% SO 2 Control Cost Summary Control Technology Control Eff % Controlled Emissions T/y Emission Reduction T/yr Installed Capital Cost $ Total Annual Cost $/yr Pollution Control Cost $/ton Incremental Control Cost $/ton Wet Walled Electrostatic Precipitator (WWESP) Line 1 Pellet Induration 80% $29,805,809 $7,441,650 $671,145 NA Line 2 Pellet Induration 80% $41,621,612 $17,974,280 $8,171 $22,600 Wet Scrubber (Additional) Line 1 Pellet Induration 60% $20,773,230 $3,196,687 $384,402 NA Line 2 Pellet Induration 60% $28,067,807 $5,545,472 $3,361 NA Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls Cost Summary 9/7/2006 Page 1 of 40

75 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 2>: - Summary of Utility, Chemical and Supply Costs Operating Unit: Line 1 Pellet Induration Study Year 2006 Emission Unit Number EU 040 Stack/Vent Number SV 046 Operating Unit: Line 2 Pellet Induration Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Line 1 furnacemmbtu 190 Line 2 furnace mmbtu 400 Reference Item Unit Cost Units Cost Year Data Source Notes Operating Labor $/hr 2006 Site Specific - United Taconite. Maintenance Labor $/hr 2006 Site Specific - United Taconite. Construction Labor Rate. Electricity 0.04 $/kwh 2006 Site Specific - United Taconite. Purchased Cost. Natural Gas $/mscf 2006 Site Specific - United Taconite Btu/scf. Water 2.07 $/mgal 2006 Site Specific - United Taconite. Purchased Cost. Cooling Water 0.28 $/mgal 0.23 EPA Air Pollution Control Cost Manual, 6th 1999 ed. Section 3.1 Ch 1. Ch 1 Carbon Absorbers, 1999 $ $0.30 Avg of 22.5 and 7 yrs and 3% inflation. EPA Air Pollution Control Cost Manual 6th Ed Example problem; Dried & Filtered, Ch 1.6 '98 cost adjusted for 3% Compressed Air 0.32 $/mscf , Section 6 Chapter 1. inflation. Wastewater Disposal Neutralization 3.13 $/mgal 2006 Site Specific - United Taconite. Purchased Cost. Chemicals & Supplies Lime $/ton Estimate from Cutler Magney Company Oxygen $/ton 2006 BOC estimate. EPA Air Pollution Control Cost Manual 6th Ed Annual costs for a retrofit SCR system example problem. '00 costs Ammonia (29% aqua.) 0.12 $/lb , Section 5 Chapter 2, page adjusted for 3% inflation. Caustic $/ton United Taconite 2006 cost Other Sales Tax 6.5% % Interest Rate 7.00% % EPA Air Pollution Control Cost Manual Introduction, Chapter 2, section Social (discount) rate used as a default. Operating Information Annual Op. Hrs EU Hours BART spreadsheet. Maximum hours of operation with a 10% safety factor. Annual Op. Hrs EU Hours BART spreadsheet. Maximum hours of operation with a 10% safety factor. Utilization Rate 93% Equipment Life 20 yrs Standardized Flow Rate SV ,345 32º F Calculated. SV 048 & ,086 32º F Calculated. Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls Utility Chem$ Data 9/7/2006 Page 2 of 40

76 Line 2 furnace mmbtu 400 Reference Item Unit Cost Units Cost Year Data Source Notes Temperature SV Deg F BART spreadsheet. SV 048 & Deg F BART spreadsheet. Weighted average. Moisture Content SV % 2005 Y:\23\69\948 United Taconite Post EERP Testing\Data\Waste Gas Stack. Average from test results in : Line 1 WG pm SV 048 & % 2005 Y:\23\69\872 United Tac\Data. Weighted average from test results in: SV049 2BPM and SV048 2APM. Actual Flow Rate SV ,137 acfm BART spreadsheet. SV 048 & ,000 acfm BART spreadsheet. Standardized Flow Rate SV ,907 68º F Calculated. SV 048 & ,092 68º F Calculated. Dry Std Flow Rate SV ,110 68º F Calculated. SV 048 & ,896 68º F Calculated. Max Emis Actual Emissions lb/hr ton/year Pollutant Lb/Hr Ton/year ('04&'05 Max) ppmv ppmv lb/mmbtu Nitrous Oxides (NOx) SV , plus 10% SV 048 & , plus 10% SV 048 & 049 with NG only , Sulfur Dioxides (SO2) SV plus 10% SV 048 & , plus 10% estimate calculated value known or assumed data required data Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls Utility Chem$ Data 9/7/2006 Page 3 of 40

77 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 3>: SO 2 Control - Wet Walled Electrostatic Precipitator Operating Unit: Line 1 Pellet Induration Emission Unit Number EU 040 Stack/Vent Number SV 046 Standardized Flow Rate 308,345 32º F Expected Utilization Rate 93% Temperature 129 Deg F Expected Annual Hours of Operation 6,599 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 369,137 acfm Expected Equipment Life 20 yrs Standardized Flow Rate 330,907 68º F Dry Std Flow Rate 285,110 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs Purchased Equipment (A) 5,420,940 Purchased Equipment Total (B) 22% of control device cost (A) 6,586,442 Installation - Standard Costs 67% of purchased equip cost (B) 4,412,916 Installation - Site Specific Costs 6,200,000 Installation Total 4,412,916 Total Direct Capital Cost, DC 10,999,359 Total Indirect Capital Costs, IC 57% of purchased equip cost (B) 3,754,272 Total Capital Investment (TCI) = DC + IC 29,805,809 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 3,935,246 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 3,506,404 Total Annual Cost (Annualized Capital Cost + Operating Cost) 7,441,650 Actual Emission Control Cost Calculation Emissions Max Emis Annual Cont Eff Cont Emis Reduction Cont Cost Pollutant Lb/Hr T/Yr % T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) , % NA Sulfur Dioxide (SO 2) % ,145 Notes & Assumptions 1 Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 6 Chapter 3. 2 Original estimate from Durr. Used 0.6 power law factor to adjust price to stack flow rate acfm from bid basis of 291,000 acfm. 3 CUECost Workbook Version 1.0, USEPA Document Page 2. Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 WWESP 9/7/2006 Page 4 of 40

78 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 3>: SO2 Control - Wet Walled Electrostatic Precipitator CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) 1 Purchased Equipment Costs (A) 2 - ESP + auxiliary equipment 5,420,940 Instrumentation 10% of control device cost (A) 542,094 MN Sales Taxes 6.5% of control device cost (A) 352,361 Freight 5% of control device cost (A) 271,047 Purchased Equipment Total (B) 22% 6,586,442 Installation Foundations & supports 4% of purchased equip cost (B) 263,458 Handling & erection 50% of purchased equip cost (B) 3,293,221 Electrical 8% of purchased equip cost (B) 526,915 Piping 1% of purchased equip cost (B) 65,864 Insulation 2% of purchased equip cost (B) 131,729 Painting 2% of purchased equip cost (B) 131,729 Installation Subtotal Standard Expenses 67% 4,412,916 Total Direct Capital Cost, DC 10,999,359 Indirect Capital Costs Engineering, supervision 20% of purchased equip cost (B) 1,317,288 Construction & field expenses 20% of purchased equip cost (B) 1,317,288 Contractor fees 10% of purchased equip cost (B) 658,644 Start-up 1% of purchased equip cost (B) 65,864 Performance test 1% of purchased equip cost (B) 65,864 Model Studies 2% of purchased equip cost (B) 131,729 Contingencies 3% of purchased equip cost (B) 197,593 Total Indirect Capital Costs, IC 57% of purchased equip cost (B) 3,754,272 Total Capital Investment (TCI) = DC + IC 14,753,631 Retrofit multiplier 3 60% of TCI 8,852,178 Sitework and foundations Site Specific 1,400,000 Structural steel Site Specific 4,800,000 Site Specific - Other Site Specific NA Total Site Specific Costs 6,200,000 Total Capital Investment (TCI) Retrofit Installed 29,805,809 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 2.0 hr/8 hr shift, Annual Operating Hours 74,749 Supervisor 15% 15% of Operator Costs 11,212 Maintenance Maintenance Labor $/Hr, 15.0 hr/wk, Annual Operating Hours 31,165 Maintenance Materials 1.00 % of Maintenance Labor 54,209 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 872 kw-hr, annual operating hours, 93% utilization 214,127 Water 2.07 $/mgal, 1,846 gpm, annual operating hours, 93% utilization 1,406,804 WW Treat Neutralization 3.13 $/mgal, 1,846 gpm, annual operating hours, 93% utilization 2,127,197 Caustic $/ton, 11 lb/hr, hr/yr, 93% utilization 15,782 Total Annual Direct Operating Costs 3,935,246 Indirect Operating Costs Overhead 60% of total labor and material costs 102,802 Administration (2% total capital costs) 2% of total capital costs (TCI) 295,073 Property tax (1% total capital costs) 1% of total capital costs (TCI) 147,536 Insurance (1% total capital costs) 1% of total capital costs (TCI) 147,536 Capital Recovery for a 20- year equipment life and a 7% interest rate 2,813,458 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 3,506,404 Total Annual Cost (Annualized Capital Cost + Operating Cost) 7,441,650 See summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 WWESP 9/7/2006 Page 5 of 40

79 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 3>: SO2 Control - Wet Walled Electrostatic Precipitator Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit Amount Required 0 Total Rep Parts Cost 0 Installation Labor 0 Total Installed Cost 0 Annualized Cost 0 Electrical Use Flow acfm D P in H2O kw Fan Power 369, EPA Cont Cost Manual 6th ed - Sec 6 Ch 3 Eq 3.46 Fluid Head (ft) Pump Eff. Pump Power 60 60% 34.8 EPA Cont Cost Manual 6th ed - Sec 6 Ch 3 Eq 3.47 Plate Area ESP Power 87, EPA Cont Cost Manual 6th ed - Sec 6 Ch 3 Eq 3.48 Total Reagent Use & Other Operating Costs gpm Water 5.00 gal/min-kacfm EPA Cont Cost Manual 6th ed - Sec 6 Ch lb/hr Reagent Use 2.50 lb NaOH/lb SO lb/hr caustic Operating Cost Calculations Annual hours of operation: 6,599 Utilization Rate: 93.0% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 2.0 hr/8 hr shift 1,650 74,749 $/Hr, 2.0 hr/8 hr shift, Annual Operating Hours Supervisor 15% of Op. NA 11,212 15% of Operator Costs Maintenance Maint Labor $/Hr 15.0 hr/wk ,165 $/Hr, 15.0 hr/wk, Annual Operating Hours Maint Mtls 1 % of Purchase Cost NA 54,209 EPA Cont Cost Manual 6th ed - Sec 6 Ch Utilities, Supplies, Replacements & Waste Management Electricity $/kwh kw-hr 5,353, ,127 $/kwh, 872 kw-hr, annual operating hours, 93% utilization Water 2.07 $/mgal 1,845.7 gpm 679,616 1,406,804 $/mgal, 1,846 gpm, annual operating hours, 93% utilization WW Treat Neutralization 3.13 $/mgal 1,845.7 gpm 679,616 2,127,197 $/mgal, 1,846 gpm, annual operating hours, 93% utilization Caustic $/ton 10.6 lb/hr 33 15,782 $/ton, 11 lb/hr, hr/yr, 93% utilization See summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 WWESP 9/7/2006 Page 6 of 40

80 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 4>: SO 2 Control - Wet Walled Electrostatic Precipitator Operating Unit: Line 2 Pellet Induration Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Standardized Flow Rate 615,086 32º F Expected Utilization Rate 93% Temperature 121 Deg F Expected Annual Hours of Operation 8,688 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 726,000 acfm Expected Equipment Life 20 yrs Standardized Flow Rate 660,092 68º F Dry Std Flow Rate 568,896 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs Purchased Equipment (A) 8,134,372 Purchased Equipment Total (B) 22% of control device cost (A) 9,883,262 Installation - Standard Costs 67% of purchased equip cost (B) 6,621,786 Installation - Site Specific Costs 6,200,000 Installation Total 6,621,786 Total Direct Capital Cost, DC 16,505,048 Total Indirect Capital Costs, IC 57% of purchased equip cost (B) 5,633,459 Total Capital Investment (TCI) = DC + IC 41,621,612 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 13,024,545 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 4,949,735 Total Annual Cost (Annualized Capital Cost + Operating Cost) 17,974,280 Actual Emission Control Cost Calculation Emissions Max Emis Annual Cont Eff Cont Emis Reduction Cont Cost Pollutant Lb/Hr T/Yr % T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) , % NA Sulfur Dioxide (SO 2) , % , ,171 Notes & Assumptions 1 Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 6 Chapter 3. 2 Original estimate from Durr. Used 0.6 power law factor to adjust price to stack flow rate acfm from bid basis of 291,000 acfm. 3 CUECost Workbook Version 1.0, USEPA Document Page 2. Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 WWESP 9/7/2006 Page 7 of 40

81 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 4>: SO2 Control - Wet Walled Electrostatic Precipitator CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) 1 Purchased Equipment Costs (A) 2 - ESP + auxiliary equipment 8,134,372 Instrumentation 10% of control device cost (A) 813,437 MN Sales Taxes 6.5% of control device cost (A) 528,734 Freight 5% of control device cost (A) 406,719 Purchased Equipment Total (B) 22% 9,883,262 Installation Foundations & supports 4% of purchased equip cost (B) 395,330 Handling & erection 50% of purchased equip cost (B) 4,941,631 Electrical 8% of purchased equip cost (B) 790,661 Piping 1% of purchased equip cost (B) 98,833 Insulation 2% of purchased equip cost (B) 197,665 Painting 2% of purchased equip cost (B) 197,665 Installation Subtotal Standard Expenses 67% 6,621,786 Total Direct Capital Cost, DC 16,505,048 Indirect Capital Costs Engineering, supervision 20% of purchased equip cost (B) 1,976,652 Construction & field expenses 20% of purchased equip cost (B) 1,976,652 Contractor fees 10% of purchased equip cost (B) 988,326 Start-up 1% of purchased equip cost (B) 98,833 Performance test 1% of purchased equip cost (B) 98,833 Model Studies 2% of purchased equip cost (B) 197,665 Contingencies 3% of purchased equip cost (B) 296,498 Total Indirect Capital Costs, IC 57% of purchased equip cost (B) 5,633,459 Total Capital Investment (TCI) = DC + IC 22,138,507 Retrofit multiplier 3 60% of TCI 13,283,104 Sitework and foundations Site Specific 1,400,000 Structural steel Site Specific 4,800,000 Site Specific - Other Site Specific NA Total Site Specific Costs 6,200,000 Total Capital Investment (TCI) Retrofit Installed 41,621,612 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 2.0 hr/8 hr shift, Annual Operating Hours 98,411 Supervisor 15% 15% of Operator Costs 14,762 Maintenance Maintenance Labor $/Hr, 15.0 hr/wk, Annual Operating Hours 31,165 Maintenance Materials 1.00 % of Maintenance Labor 81,344 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 1,716 kw-hr, annual operating hours, 93% utilization 554,444 Water 2.07 $/mgal, 3,630 gpm, annual operating hours, 93% utilization 3,642,680 WW Treat Neutralization 3.13 $/mgal, 3,630 gpm, annual operating hours, 93% utilization 5,508,013 Caustic $/ton, 1,580 lb/hr, hr/yr, 93% utilization 3,093,727 Total Annual Direct Operating Costs 13,024,545 Indirect Operating Costs Overhead 60% of total labor and material costs 135,409 Administration (2% total capital costs) 2% of total capital costs (TCI) 442,770 Property tax (1% total capital costs) 1% of total capital costs (TCI) 221,385 Insurance (1% total capital costs) 1% of total capital costs (TCI) 221,385 Capital Recovery for a 20- year equipment life and a 7% interest rate 3,928,786 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 4,949,735 Total Annual Cost (Annualized Capital Cost + Operating Cost) 17,974,280 See summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 WWESP 9/7/2006 Page 8 of 40

82 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 4>: SO2 Control - Wet Walled Electrostatic Precipitator Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit Amount Required 0 Total Rep Parts Cost 0 Installation Labor 0 Total Installed Cost 0 Annualized Cost 0 Electrical Use Flow acfm D P in H2O kw Fan Power 726, ,314.1 EPA Cont Cost Manual 6th ed - Sec 6 Ch 3 Eq 3.46 Fluid Head (ft) Pump Eff. Pump Power 60 60% 68.4 EPA Cont Cost Manual 6th ed - Sec 6 Ch 3 Eq 3.47 Plate Area ESP Power 171, EPA Cont Cost Manual 6th ed - Sec 6 Ch 3 Eq 3.48 Total 1,715.6 Reagent Use & Other Operating Costs gpm Water 5.00 gal/min-kacfm EPA Cont Cost Manual 6th ed - Sec 6 Ch lb/hr Reagent Use 2.50 lb NaOH/lb SO lb/hr caustic Operating Cost Calculations Annual hours of operation: 8,688 Utilization Rate: 93.0% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 2.0 hr/8 hr shift 2,172 98,411 $/Hr, 2.0 hr/8 hr shift, Annual Operating Hours Supervisor 15% of Op. NA 14,762 15% of Operator Costs Maintenance Maint Labor $/Hr 15.0 hr/wk ,165 $/Hr, 15.0 hr/wk, Annual Operating Hours Maint Mtls 1 % of Purchase Cost NA 81,344 EPA Cont Cost Manual 6th ed - Sec 6 Ch Utilities, Supplies, Replacements & Waste Management Electricity $/kwh kw-hr 13,861, ,444 $/kwh, 1,716 kw-hr, annual operating hours, 93% utilization Water 2.07 $/mgal 3,630.0 gpm 1,759,749 3,642,680 $/mgal, 3,630 gpm, annual operating hours, 93% utilization WW Treat Neutralization 3.13 $/mgal 3,630.0 gpm 1,759,749 5,508,013 $/mgal, 3,630 gpm, annual operating hours, 93% utilization Caustic $/ton 1,580.4 lb/hr 6,385 3,093,727 $/ton, 1,580 lb/hr, hr/yr, 93% utilization See summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 WWESP 9/7/2006 Page 9 of 40

83 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 5>: SO x Control - Wet Scrubber Operating Unit: Line 1 Pellet Induration Emission Unit Number EU 040 Stack/Vent Number SV 046 Standardized Flow Rate 308,345 32º F Expected Utilization Rate 93% Temperature 129 Deg F Expected Annual Hours of Operation 6,599 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 369,137 acfm Expected Equipment Life 20 yrs Standardized Flow Rate 330,907 68º F Dry Std Flow Rate 285,110 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs (1) Purchased Equipment (A) 3,748,259 Purchased Equipment Total (B) 22% of control device cost (A) 4,554,134 Installation - Standard Costs 85% of purchased equip cost (B) 3,871,014 Installation - Site Specific Costs 6,200,000 Installation Total 3,871,014 Total Direct Capital Cost, DC 8,425,148 Total Indirect Capital Costs, IC 15% of purchased equip cost (B) 683,120 Total Capital Investment (TCI) = DC + IC 20,773,230 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 835,246 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 2,361,441 Total Annual Cost (Annualized Capital Cost + Operating Cost) 3,196,687 Actual Emission Control Cost Calculation Emissions Max Emis Annual Cont Eff Cont Emis Reduction Cont Cost Pollutant Lb/Hr T/Yr % T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) , % NA Sulfur Dioxide (SO 2 ) % ,402 Notes & Assumptions 1 Original estimate from STS Consultants. Use 0.6 Power law factor to adjust price to stack flow rate from bid basis of 500,000 acfm. 2 Liquid/Gas ratio = 38 L/G = Gal/1,000 acf. 3 Water Makeup Rate/Wastewater Discharge = 2.0% of circulating water rate. 4 Evaporation rate calculated from steam table in Basic Principles and Calculations in Chemical Engineering Third Edition. 5 CUECost Workbook Version 1.0, USEPA Document Page 2. Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Wet Scrubber 9/7/2006 Page 10 of 40

84 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 5>: SOx Control - Wet Scrubber CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) Purchased Equipment Costs (A) - Absorber + packing + auxiliary equipment, EC 3,748,259 Instrumentation 10% of control device cost (A) 374,826 MN Sales Taxes 7% of control device cost (A) 243,637 Freight 5% of control device cost (A) 187,413 Purchased Equipment Total (B) 22% 4,554,134 Installation Foundations & supports 12% of purchased equip cost (B) 546,496 Handling & erection 40% of purchased equip cost (B) 1,821,654 Electrical 1% of purchased equip cost (B) 45,541 Piping 30% of purchased equip cost (B) 1,366,240 Insulation 1% of purchased equip cost (B) 45,541 Painting 1% of purchased equip cost (B) 45,541 Installation Subtotal Standard Expenses 85% 3,871,014 Total Direct Capital Cost, DC 8,425,148 Indirect Capital Costs Engineering, supervision 5% of purchased equip cost (B) 227,707 Construction & field expenses 0% of purchased equip cost (B) 0 Contractor fees 5% of purchased equip cost (B) 227,707 Start-up 1% of purchased equip cost (B) 45,541 Performance test 1% of purchased equip cost (B) 45,541 Model Studies NA of purchased equip cost (B) NA Contingencies 3% of purchased equip cost (B) 136,624 Total Indirect Capital Costs, IC 15% of purchased equip cost (B) 683,120 Total Capital Investment (TCI) = DC + IC 9,108,269 Retrofit multiplier 5 60% of TCI 5,464,961 Sitework and foundations Site Specific 1,400,000 Structural steel Site Specific 4,800,000 Site Specific - Other Site Specific 0 Total Site Specific Costs 6,200,000 Total Capital Investment (TCI) Retrofit Installed 20,773,230 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Wet Scrubber 9/7/2006 Page 11 of 40

85 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 5>: SOx Control - Wet Scrubber OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours 18,687 Supervisor 15% 15% of Operator Costs 2,803 Maintenance Maintenance Labor $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours 19,475 Maintenance Materials 100% of maintenance labor costs 19,475 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 759 kw-hr, annual operating hours, 93% utilization 186,351 Water 2.07 $/kgal, 346 gpm, annual operating hours, 93% utilization 264,005 WW Treat Neutralization 3.13 $/kgal, 281 gpm, annual operating hours, 93% utilization 323,334 Lime $/ton, 4 lb/hr, annual operating hours, 93% utilization 1,116 Total Annual Direct Operating Costs 835,246 Indirect Operating Costs Overhead 60% of total labor and material costs 36,264 Administration (2% total capital costs) 2% of total capital costs (TCI) 182,165 Property tax (1% total capital costs) 1% of total capital costs (TCI) 91,083 Insurance (1% total capital costs) 1% of total capital costs (TCI) 91,083 Capital Recovery for a 20- year equipment life and a 7% interest rate 1,960,846 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 2,361,441 Total Annual Cost (Annualized Capital Cost + Operating Cost) 3,196,687 See Summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Wet Scrubber 9/7/2006 Page 12 of 40

86 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 5>: SOx Control - Wet Scrubber Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Catalyst: Equipment Life 20 years CRF Rep part cost per unit 0 $/ft 3 Amount Required 0 ft 3 Packing Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 Assume Labor = 15% of catalyst cost (basis labor for baghouse replacement) Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit 0.00 $ each Amount Required 0 Number Total Rep Parts Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 10 min per bag (13 hr total) Labor at $29.65/hr OAQPS list replacement times from 5-20 min per bag. Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Electrical Use Flow acfm P in H2O Efficiency Hp kw Blower, Scrubber 369, EPA Cost Cont Manual 6th ed Section 5.2 Chapter 1 Eq 1.48 Flow Liquid SPGR P ft H2O Efficiency Hp kw Circ Pump 14,027 gpm EPA Cost Cont Manual 6th ed Section 5.2 Chapter 1 Eq 1.49 H2O WW Disch 346 gpm EPA Cost Cont Manual 6th ed Section 5.2 Chapter 1 Eq 1.49 Other Total Reagent Use & Other Operating Costs Caustic Use 4.25 lb/hr SO lb NaOH/lb SO lb/hr Caustic Lime Use 4.25 lb/hr SO lb Lime/lb SO lb/hr lime, lime addition at 1.1 times the stoichiometric ratio Liquid/Gas ratio 38.0 * L/G = Gal/1,000 acf Circulating Water Rate 2 14,027 gpm Water Makeup Rate/WW Disch 3 = 2.0% of circulating water rate + evap. loss = 346 gpm Evaporation Loss 4 = gpm Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Wet Scrubber 9/7/2006 Page 13 of 40

87 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 5>: SOx Control - Wet Scrubber Operating Cost Calculations Annual hours of operation: 6,599 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.5 hr/8 hr shift ,687 $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours Supervisor 15% of Op. NA 2,803 15% of Operator Costs Maintenance Maint Labor $/Hr 0.5 hr/8 hr shift ,475 $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours Maint Mtls 100 % of Maintenance Labor NA 19, % of Maintenance Labor Utilities, Supplies, Replacements & Waste Management Electricity $/kwh kw-hr 4,658, ,351 $/kwh, 759 kw-hr, annual operating hours, 93% utilization Water 2.07 $/kgal gpm 127, ,005 $/kgal, 346 gpm, annual operating hours, 93% utilization WW Treat Neutralization 3.13 $/kgal gpm 103, ,334 $/kgal, 281 gpm, annual operating hours, 93% utilization Lime 89.0 $/ton 4.1 lb/hr 13 1,116 $/ton, 4 lb/hr, annual operating hours, 93% utilization See Summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Wet Scrubber 9/7/2006 Page 14 of 40

88 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 6>: SO x Control - Wet Scrubber Operating Unit: Line 2 Pellet Induration Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Standardized Flow Rate 615,086 32º F Expected Utilization Rate 93% Temperature 121 Deg F Expected Annual Hours of Operation 8,688 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 726,000 acfm Expected Equipment Life 20 yrs Standardized Flow Rate 660,092 68º F Dry Std Flow Rate 568,896 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs (1) Purchased Equipment (A) 5,624,436 Purchased Equipment Total (B) 22% of control device cost (A) 6,833,690 Installation - Standard Costs 85% of purchased equip cost (B) 5,808,636 Installation - Site Specific Costs 6,200,000 Installation Total 5,808,636 Total Direct Capital Cost, DC 12,642,326 Total Indirect Capital Costs, IC 15% of purchased equip cost (B) 1,025,053 Total Capital Investment (TCI) = DC + IC 28,067,807 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 2,301,630 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 3,243,841 Total Annual Cost (Annualized Capital Cost + Operating Cost) 5,545,472 Actual Emission Control Cost Calculation Emissions Max Emis Annual Cont Eff Cont Emis Reduction Cont Cost Pollutant Lb/Hr T/Yr % T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) , % NA Sulfur Dioxide (SO 2 ) , % , ,361 Notes & Assumptions 1 Original estimate from STS Consultants. Use 0.6 Power law factor to adjust price to stack flow rate from bid basis of 500,000 acfm. 2 Liquid/Gas ratio = 38 L/G = Gal/1,000 acf. 3 Water Makeup Rate/Wastewater Discharge = 2.0% of circulating water rate. 4 Evaporation rate calculated from steam table in Basic Principles and Calculations in Chemical Engineering Third Edition. 5 CUECost Workbook Version 1.0, USEPA Document Page 2. Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Wet Scrubber 9/7/2006 Page 15 of 40

89 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 6>: SOx Control - Wet Scrubber CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) Purchased Equipment Costs (A) - Absorber + packing + auxiliary equipment, EC 5,624,436 Instrumentation 10% of control device cost (A) 562,444 MN Sales Taxes 7% of control device cost (A) 365,588 Freight 5% of control device cost (A) 281,222 Purchased Equipment Total (B) 22% 6,833,690 Installation Foundations & supports 12% of purchased equip cost (B) 820,043 Handling & erection 40% of purchased equip cost (B) 2,733,476 Electrical 1% of purchased equip cost (B) 68,337 Piping 30% of purchased equip cost (B) 2,050,107 Insulation 1% of purchased equip cost (B) 68,337 Painting 1% of purchased equip cost (B) 68,337 Installation Subtotal Standard Expenses 85% 5,808,636 Total Direct Capital Cost, DC 12,642,326 Indirect Capital Costs Engineering, supervision 5% of purchased equip cost (B) 341,684 Construction & field expenses 0% of purchased equip cost (B) 0 Contractor fees 5% of purchased equip cost (B) 341,684 Start-up 1% of purchased equip cost (B) 68,337 Performance test 1% of purchased equip cost (B) 68,337 Model Studies NA of purchased equip cost (B) NA Contingencies 3% of purchased equip cost (B) 205,011 Total Indirect Capital Costs, IC 15% of purchased equip cost (B) 1,025,053 Total Capital Investment (TCI) = DC + IC 13,667,380 Retrofit multiplier 5 60% of TCI 8,200,428 Sitework and foundations Site Specific 1,400,000 Structural steel Site Specific 4,800,000 Site Specific - Other Site Specific 0 Total Site Specific Costs 6,200,000 Total Capital Investment (TCI) Retrofit Installed 28,067,807 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Wet Scrubber 9/7/2006 Page 16 of 40

90 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 6>: SOx Control - Wet Scrubber OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours 24,603 Supervisor 15% 15% of Operator Costs 3,690 Maintenance Maintenance Labor $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours 25,640 Maintenance Materials 100% of maintenance labor costs 25,640 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 1,493 kw-hr, annual operating hours, 93% utilizatio 482,525 Water 2.07 $/kgal, 681 gpm, annual operating hours, 93% utilization 683,595 WW Treat Neutralization 3.13 $/kgal, 552 gpm, annual operating hours, 93% utilization 837,218 Lime $/ton, 608 lb/hr, annual operating hours, 93% utilization 218,719 Total Annual Direct Operating Costs 2,301,630 Indirect Operating Costs Overhead 60% of total labor and material costs 47,744 Administration (2% total capital costs) 2% of total capital costs (TCI) 273,348 Property tax (1% total capital costs) 1% of total capital costs (TCI) 136,674 Insurance (1% total capital costs) 1% of total capital costs (TCI) 136,674 Capital Recovery for a 20- year equipment life and a 7% interest rate 2,649,402 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 3,243,841 Total Annual Cost (Annualized Capital Cost + Operating Cost) 5,545,472 See Summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Wet Scrubber 9/7/2006 Page 17 of 40

91 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 6>: SOx Control - Wet Scrubber Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Catalyst: Equipment Life 20 years CRF Rep part cost per unit 0 $/ft 3 Amount Required 0 ft 3 Packing Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 Assume Labor = 15% of catalyst cost (basis labor for baghouse replacement) Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit 0.00 $ each Amount Required 0 Number Total Rep Parts Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 10 min per bag (13 hr total) Labor at $29.65/hr OAQPS list replacement times from 5-20 min per bag. Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Electrical Use Flow acfm P in H2O Efficiency Hp kw Blower, Scrubber 726, ,037.5 EPA Cost Cont Manual 6th ed Section 5.2 Chapter 1 Eq 1.48 Flow Liquid SPGR P ft H2O Efficiency Hp kw Circ Pump 27,588 gpm EPA Cost Cont Manual 6th ed Section 5.2 Chapter 1 Eq 1.49 H2O WW Disch 681 gpm EPA Cost Cont Manual 6th ed Section 5.2 Chapter 1 Eq 1.49 Other Total Reagent Use & Other Operating Costs Caustic Use lb/hr SO lb NaOH/lb SO lb/hr Caustic Lime Use lb/hr SO lb Lime/lb SO lb/hr lime, lime addition at 1.1 times the stoichiometric ratio Liquid/Gas ratio 38.0 * L/G = Gal/1,000 acf Circulating Water Rate 2 27,588 gpm Water Makeup Rate/WW Disch 3 = 2.0% of circulating water rate + evap. loss = 681 gpm Evaporation Loss 4 = gpm Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Wet Scrubber 9/7/2006 Page 18 of 40

92 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 6>: SOx Control - Wet Scrubber Operating Cost Calculations Annual hours of operation: 8,688 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.5 hr/8 hr shift ,603 $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours Supervisor 15% of Op. NA 3,690 15% of Operator Costs Maintenance Maint Labor $/Hr 0.5 hr/8 hr shift ,640 $/Hr, 0.5 hr/8 hr shift, Annual Operating Hours Maint Mtls 100 % of Maintenance Labor NA 25, % of Maintenance Labor Utilities, Supplies, Replacements & Waste Management Electricity $/kwh kw-hr 12,063, ,525 $/kwh, 1,493 kw-hr, annual operating hours, 93% utilization Water 2.07 $/kgal gpm 330, ,595 $/kgal, 681 gpm, annual operating hours, 93% utilization WW Treat Neutralization 3.13 $/kgal gpm 267, ,218 $/kgal, 552 gpm, annual operating hours, 93% utilization Lime 89.0 $/ton lb/hr 2, ,719 $/ton, 608 lb/hr, annual operating hours, 93% utilization See Summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Wet Scrubber 9/7/2006 Page 19 of 40

93 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 7: -Summary of Utility, Chemical and Supply Costs Operating Unit: Line 1 Pellet Induration Emission Unit Number EU 040 Stack/Vent Number SV 046 Desgin Capacity 190 MMBtu/hr Standardized Flow Rate 308,345 32º F Expected Utiliztion Rate 93% Temperature 129 Deg F Expected Annual Hours of Operation 6,599 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 369,137 acfm Expected Equipment Life 20 yrs Standardized Flow Rate 330,907 68º F Dry Std Flow Rate 285,110 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs EPRI Correlation Purchased Equipment (A) 15,894,553 Purchased Equipment Total (B) SCR Only 16,927,699 Installation - Standard Costs 15% of purchased equip cost (B) SCR Only 2,935,740 Installation - Site Specific Costs 0 Installation Total 0 Total Direct Capital Cost, DC 0 Total Indirect Capital Costs, IC 0% of purchased equip cost (B) 0 Total Capital Investment (TCI) = DC + IC SCR + Reheat 44,591,384 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. SCR + Reheat 18,389,200 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost SCR + Reheat 5,534,036 Total Annual Cost (Annualized Capital Cost + Operating Cost) SCR + Reheat 23,923,236 Emission Control Cost Calculation Baseline Predicted Limit Cont. Emis. Cont Emis Reduction Cont Cost Pollutant Emis. T/yr lb/hr lb/mmbtu T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) 2, , ,659 Notes & Assumptions 1 Estimated Equipment Cost per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2. 2 For Calculation purposes, duty reflects increased flow rate, not actual duty. 3 Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 4 Capital Cost per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq Reagent Use per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Catalyst Volume per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Reactor Size per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Catalyst Replacement per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Electrical Demand per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Maintenance Costs EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq Control Efficiency = 80% reduction which is the normal SCR control efficiency 12 $35/MW-hr, 140 MW 13 Catalyst replacement every three years. This requires an additional 2 week outage per 3 year outage cycle, annualized to 4.7 days. Notes to User Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 SCR EU040 SCR 20 of 40

94 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 7: -Summary of Utility, Chemical and Supply Costs CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) Purchased Equipment Costs (A) - Absorber + packing + auxillary equipment, EC 15,894,553 Instrumentation 10% of control device cost (A) NA MN Sales Taxes 6.5% of control device cost (A) 1,033,146 Freight 5% of control device cost (A) NA Purchased Equipment Total (A) 16,927,699 Indirect Installation General Facilities 0% of purchased equip cost (A) 0 Engineering & Home Office 0% of purchased equip cost (A) 0 Process Contingency 0% of purchased equip cost (A) 0 Site Specific-Other 16% Replacement Power, two weeks 2,643,899 Total Indirect Installation Costs (B) 16% of purchased equip cost (A) 2,643,899 Project Contingeny ( C) 15% of (A + B) 2,935,740 Total Plant Cost D A + B + C 22,507,337 Allowance for Funds During Construction (E) Pre Production Costs (G) 2% of (D+E)) 450,147 Inventory Capital Reagent Vol * $/gal 289,894 Total Capital Investment (TCI) = DC + IC D + E + F + G +H + I 23,247,378 Retrofit multiplier 3 60% of TCI 13,948,427 Adjusted TCI for Replacement Parts (Catalyst, Filter Bags, etc) for Capital Recovery Cost NA Total Retrofit Capital 37,195,804 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 2.0 hr/8 hr shift, hr/yr 74,749 Supervisor 15% 15% of Operator Costs 11,212 Maintenance Maintenance Total 1.50 % of Total Capital Investment 348,711 Maintenance Materials NA % of Maintenance Labor - Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 5,919 kw-hr, hr/yr, 93% utilization 1,453,059 NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - Cat. Replacement [14] Catalyst Replacement 747,654 NA NA - Ammonia 0.12 $/lb, 8,542 lb/hr, hr/yr, 93% utilization 6,322,342 NA NA - NA NA - NA NA - Total Annual Direct Operating Costs 8,957,727 Indirect Operating Costs Overhead 60% fixed O&M 62,942 Administration (2% total capital costs) 2% of total capital costs (TCI) 464,948 Property tax (1% total capital costs) 1% of total capital costs (TCI) 232,474 Insurance (1% total capital costs) 1% of total capital costs (TCI) 232,474 Capital Recovery 9.44% for a 20- year equipment life and a 7% interest rate 3,511,021 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 4,503,858 Total Annual Cost (Annualized Capital Cost + Operating Cost) 13,461,585 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 SCR EU040 SCR 21 of 40

95 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 7: -Summary of Utility, Chemical and Supply Costs Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Equivalent Duty 1,859 Plant Cap kw A 190,728 Est power platn eff 35% Unc Nox lb/mmbtu B 3.43 D=75*(300,000)*((B/1.5)^0.05*(C/100)^0.04)A)^0.35 Watt per Btu/hr Nox Red. Eff. C 80% E=D*A* Equivalent power plant k 190,728 Capital Cost $/kw D $83.34 $15,894, Total SCR Equipment Uncontrolled Nox t/y 2,151.2 Fixed O&M E $104, Annual Operating Hrs 8000 Variable O&M F $423, Uncontrolled Nox lb/mm Ann Cap Factor G 0.82 Heat Input mmbtu/ H 6,000 SCR Capital Cost Duty 1,859 MMBtu/hr Catalyst Area 897 ft 2 1,352 f (h SCR) Q flue gas 861,164 acfm Rx Area 1, f (h NH 3 ) NOx Cont Eff 80% (as faction) Rx Height 32.1 ft 0 f (h New) new= -728, Retrofit = 0 NOx in 3.43 lb/mmbtu n layer 51 layers Y Bypass? Y or N Ammonia Slip 2 ppm h layer 53.6 ft 127 f (h Bypass) Fuel Sulfur 0.67 wt % (as %) n total 52 layers 34,003,097 f (vol catalyst) Temperature 129 Deg F h SCR 252 ft f (h SCR) Catalyst Volume 141,680 ft 3 New/Retrofit R N or R Electrical Use Duty 1,859 MMBtu/hr kw NOx Cont Eff 80% (as faction) Power 5,919.3 NOx in 3.43 lb/mmbtu n catalyst layers 52 layers Press drop catalyst 1 in H 2 O per layer Press drop duct 3 in H 2 O Total Reagent Use & Other Operating Costs Ammonia Use 56.0 lb/ft 3 Density 2477 lb/hr Neat gal/hr 29% solution Volume 14 day inventory 383,433 gal $289,894 Inventory Cost 8542 lb/hr Design Basis Baseline Em Baseline Emi Max Emis. (Model) Control Eff (%) Cont. Emis (lb/mmbtu) T/yr lb/mmbtu lb/hr 80% 0.69 Nitrous Oxide 2, Actual Method 19 Factor Adjusted Duty 90,035 dscf/mmbtu 9,200 dscf/mmbtu 1,859 MMBtu/hr Operating Cost Calculations Annual hours of operation: 6,599 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 2.0 hr/8 hr shift 1,650 74,749 $/Hr, 2.0 hr/8 hr shift, hr/yr Supervisor 15% of Op. NA 11,212 15% of Operator Costs Maintenance Maintenance Total 1.5 % of Total Capital Investment 348,711 % of Total Capital Investment Maint Mtls 0 % of Maintenance Labor NA 0 0% of Maintenance Labor Utilities, Supplies, Replacements & Waste Management Electricity $/kwh kw-hr 36,326,476 1,453,059 $/kwh, 5,919 kw-hr, hr/yr, 93% utilization Cat. Replacement [14] 35 $/MW-hr mw 112 $747,654 Catalyst Replacement 7 Ammonia 0.12 $/lb 8542 lb/hr 52,424,372 6,322,342 $/lb, 8,542 lb/hr, hr/yr, 93% utilization See Summary page for notes and assumptions ** Std Air use is 2 scfm/kacfm *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 SCR EU040 SCR 22 of 40

96 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 8>: - Summary of Utility, Chemical and Supply Costs Operating Unit: Line 1 Pellet Induration Emission Unit Number EU 040 Stack/Vent Number SV 046 Chemical Engineering Desgin Capacity 190 MMBtu/hr Standardized Flow Rate 308,345 32º F Chemical Plant Cost Index Expected Utiliztion Rate 93% Temperature 129 Deg F 1998/ Expected Annual Hours of Operation 6,599 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 369,137 acfm Inflation Adj 1.19 Expected Equipment Life 20 yrs Standardized Flow Rate 330,907 68º F Dry Std Flow Rate 285,110 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs Purchased Equipment (A) 611,190 Purchased Equipment Total (B) 22% of control device cost (A) 742,596 Installation - Standard Costs 30% of purchased equip cost (B) 222,779 Installation - Site Specific Costs 6,200,000 Installation Total 6,422,779 Total Direct Capital Cost, DC 7,165,375 Total Indirect Capital Costs, IC 31% of purchased equip cost (B) 230,205 Total Capital Investment (TCI) = DC + IC 7,395,580 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 9,431,473 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 1,030,178 Total Annual Cost (Annualized Capital Cost + Operating Cost) 10,461,651 Notes & Assumptions 1 Equipment cost estimate EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 3.2 Chapter Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 3.2 Chapter 2 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Reheat EU040 Reheat 23 of 40

97 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 8>: - Summary of Utility, Chemical and Supply Costs CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) 611,190 Purchased Equipment Costs (A) - Absorber + packing + auxillary equipment, EC Instrumentation 10% of control device cost (A) 61,119 MN Sales Taxes 6.5% of control device cost (A) 39,727 Freight 5% of control device cost (A) 30,560 Purchased Equipment Total (B) 22% 742,596 Installation Foundations & supports 8% of purchased equip cost (B) 59,408 Handling & erection 14% of purchased equip cost (B) 103,963 Electrical 4% of purchased equip cost (B) 29,704 Piping 2% of purchased equip cost (B) 14,852 Insulation 1% of purchased equip cost (B) 7,426 Painting 1% of purchased equip cost (B) 7,426 Installation Subtotal Standard Expenses 30% 222,779 Site Preparation, as required Site Specific 1,400,000 Buildings, as required Site Specific 4,800,000 Site Specific - Other Site Specific 0 Total Site Specific Costs 6,200,000 Installation Total 6,422,779 Total Direct Capital Cost, DC 7,165,375 Indirect Capital Costs Engineering, supervision 10% of purchased equip cost (B) 74,260 Construction & field expenses 5% of purchased equip cost (B) 37,130 Contractor fees 10% of purchased equip cost (B) 74,260 Start-up 2% of purchased equip cost (B) 14,852 Performance test 1% of purchased equip cost (B) 7,426 Model Studies of purchased equip cost (B) 0 Contingencies 3% of purchased equip cost (B) 22,278 Total Indirect Capital Costs, IC 31% of purchased equip cost (B) 230,205 Total Capital Investment (TCI) = DC + IC 7,395,580 Adjusted TCI for Replacement Parts (Catalyst, Filter Bags, etc) for Capital Recovery Cost 7,395,580 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 0.5 hr/8 hr shift, hr/yr 18,687 Supervisor 15% 15% of Operator Costs 2,803 Maintenance Maintenance Labor $/Hr, 0.5 hr/8 hr shift, hr/yr 19,475 Maintenance Materials 100% of maintenance labor costs 19,475 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 1,368 kw-hr, hr/yr, 93% utilization 335,730 Natural Gas $/mscf, 1,915 scfm, hr/yr, 93% utilization 9,035,303 Total Annual Direct Operating Costs 9,431,473 Indirect Operating Costs Overhead 60% of total labor and material costs 36,264 Administration (2% total capital costs) 2% of total capital costs (TCI) 147,912 Property tax (1% total capital costs) 1% of total capital costs (TCI) 73,956 Insurance (1% total capital costs) 1% of total capital costs (TCI) 73,956 Capital Recovery for a 20- year equipment life and a 7% interest rate 698,090 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 1,030,178 Total Annual Cost (Annualized Capital Cost + Operating Cost) 10,461,651 See Summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Reheat EU040 Reheat 24 of 40

98 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 4: -Summary of Utility, Chemical ans Supply Costs Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Catalyst: Catalyst Equipment Life 3 years CRF Rep part cost per unit 0 $/ft 3 Amount Required 39 ft 3 Catalyst Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 Assume Labor = 15% of catalyst cost (basis labor for baghouse replacement) Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit 0 $ each Amount Required 0 Number Total Rep Parts Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 10 min per bag (13 hr total) Labor at $29.65/hr OAQPS list replacement times from 5-20 min per bag. Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Electrical Use Flow acfm P in H2O Efficiency Hp kw Blower, Thermal 369, ,367.7 EPA Cost Cont Manual 6th ed - Oxidizders Chapter Blower, Catalytic 369, ,655.6 EPA Cost Cont Manual 6th ed - Oxidizders Chapter Oxidizer Type thermal (catalytic or thermal) Reagent Use & Other Operating Costs Oxidizers - NA Operating Cost Calculations Annual hours of operation: 6,599 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.5 hr/8 hr shift ,687 $/Hr, 0.5 hr/8 hr shift, hr/yr Supervisor 15% of Op. NA 2,803 15% of Operator Costs Maintenance Maint Labor $/Hr 0.5 hr/8 hr shift ,475 $/Hr, 0.5 hr/8 hr shift, hr/yr Maint Mtls 100 % of Maintenance Labor NA 19, % of Maintenance Labor Utilities, Supplies, Replacements & Waste Management Electricity $/kwh kw-hr 8,393, ,730 $/kwh, 1,368 kw-hr, hr/yr, 93% utilization Natural Gas $/mscf 1,915 scfm 705,149 9,035,303 $/mscf, 1,915 scfm, hr/yr, 93% utilization See Summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Reheat EU040 Reheat 25 of 40

99 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis <Table 8>: - Summary of Utility, Chemical and Supply Costs Flue Gas Re-Heat Equipment Cost Estimate Basis Thermal Oxidizer with 70% Heat Recovery Auxiliary Fuel Use Equation 3.19 T wi 129 Deg F - Temperature of waste gas into heat recovery T fi 750 Deg F - Temperature of Flue gas into of heat recovery T ref FER 77 Deg F - Reference temperature for fuel combustion calculations 70% Factional Heat Recovery % Heat recovery section efficiency T wo 564 Deg F - Temperature of waste gas out of heat recovery T fo 315 Deg F - Temperature of flue gas into of heat recovery -h caf Btu/lb Heat of combustion auxiliary fuel (methane) -h wg 0 Btu/lb Heat of combustion waste gas C p wg p wg p af Q wg Btu/lb - Deg F Heat Capacity of waste gas (air) lb/scf - Density of waste gas (air) at 77 Deg F lb/scf - Density of auxiliary fuel (methane) at 77 Deg F 330,907 scfm - Flow of waste gas Q af 1,915 scfm - Flow of auxiliary fuel Year 2005 Inflation Rate 3.0% Cost Calculations 332,822 scfm Flue Gas Cost in 1989 $'s $512,611 Current Cost Using CHE Plant Cost Index $611,190 Heat Rec % A B 0 10, Exponents per equation , Exponents per equation , Exponents per equation , Exponents per equation 3.27 Indurator Flue Gas Heat Capacity - Basis Typical Composition 100 scfm 359 scf/lbmole Gas Composition lb/hr f wt % Cp Gas Cp Flue 28 mw CO 0 v % 0 44 mw CO2 15 v % % mw H2O 10 v % % mw N2 60 v % % mw O2 15 v % % Cp Flue Gas 100 v % % Reference: OAQPS Control Cost Manual 5th Ed Feb Chapter 3 Thermal & Catalytic Incinerators (EPA 453/B ) Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU040 Reheat EU040 Reheat 26 of 40

100 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 9: -Summary of Utility, Chemical and Supply Costs Operating Unit: Line 2 Pellet Induration Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Chemical Engineering Desgin Capacity 400 MMBtu/hr Standardized Flow Rate 615,086 32º F Chemical Plant Cost Index Expected Utiliztion Rate 93% Temperature 121 Deg F 1998/ Expected Annual Hours of Operation 8,688 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 726,000 acfm Inflation Adj 1.19 Expected Equipment Life 20 yrs Standardized Flow Rate 660,092 68º F Dry Std Flow Rate 568,896 68º F CONTROL EQUIPMENT COSTS Capital Costs Year Direct Capital Costs EPRI Correlation 1998 Purchased Equipment (A) ,410,519 Purchased Equipment Total (B) SCR Only 23,867,203 Installation - Standard Costs 15% of purchased equip cost (B) SCR Only 3,976,665 Installation - Site Specific Costs 0 Installation Total 0 Total Direct Capital Cost, DC 0 Total Indirect Capital Costs, IC 0% of purchased equip cost (B) 0 Total Capital Investment (TCI) = DC + IC SCR + Reheat 70,762,832 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. SCR + Reheat 28,175,324 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost SCR + Reheat 7,013,819 Total Annual Cost (Annualized Capital Cost + Operating Cost) SCR + Reheat 35,189,143 Emission Control Cost Calculation Baseline Predicted Limit Cont. Emis. Cont Emis Reduction Cont Cost Pollutant Emis. T/yr lb/hr lb/mmbtu T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) 1, , ,017 Notes & Assumptions 1 Estimated Equipment Cost per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2. 2 For Calculation purposes, duty reflects increased flow rate, not actual duty. 3 Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 4 Capital Cost per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq Reagent Use per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Catalyst Volume per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Reactor Size per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Catalyst Replacement per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Electrical Demand per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Maintenance Costs EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq Control Efficiency = 80% reduction which is the normal SCR control efficiency 12 $35/MW-hr, 140 MW 13 Catalyst replacement every three years. This requires an additional 2 week outage per 3 year outage cycle, annualized to 4.7 days. Notes to User Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 SCR EU042 SCR 27 of 40

101 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 9: -Summary of Utility, Chemical and Supply Costs CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) Purchased Equipment Costs (A) - Absorber + packing + auxillary equipment, EC 22,410,519 Instrumentation 10% of control device cost (A) NA MN Sales Taxes 6.5% of control device cost (A) 1,456,684 Freight 5% of control device cost (A) NA Purchased Equipment Total (A) 23,867,203 Indirect Installation General Facilities 0% of purchased equip cost (A) 0 Engineering & Home Office 0% of purchased equip cost (A) 0 Process Contingency 0% of purchased equip cost (A) 0 Site Specific-Other 11% Replacement Power, two weeks 2,643,899 Total Indirect Installation Costs (B) 11% of purchased equip cost (A) 2,643,899 Project Contingeny ( C) 15% of (A + B) 3,976,665 Total Plant Cost D A + B + C 30,487,767 Allowance for Funds During Construction (E) Additional 10 week outage for installation 8,232,000 Pre Production Costs (G) 2% of (D+E)) 774,395 Inventory Capital Reagent Vol * $/gal 17,062 Total Capital Investment (TCI) = DC + IC D + E + F + G +H + I 39,511,225 Retrofit multiplier 3 60% of TCI 23,706,735 Adjusted TCI for Replacement Parts (Catalyst, Filter Bags, etc) for Capital Recovery Cost NA Total Retrofit Capital 63,217,960 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator NA - Supervisor NA - Maintenance Maintenance Total 1.50 % of Total Capital Investment 592,668 Maintenance Materials NA % of Maintenance Labor - Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 5,133 kw-hr, hr/yr, 93% utilization 1,658,998 NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - Cat. Replacement [14] Catalyst Replacement 548,800 NA NA - Ammonia 0.12 $/lb, 503 lb/hr, hr/yr, 93% utilization 489,909 NA NA - NA NA - NA NA - Total Annual Direct Operating Costs 3,290,375 Indirect Operating Costs Overhead NA of total labor and material costs NA Administration (2% total capital costs) NA of total capital costs (TCI) NA Property tax (1% total capital costs) NA of total capital costs (TCI) NA Insurance (1% total capital costs) NA of total capital costs (TCI) NA Capital Recovery for a 20- year equipment life and a 7% interest rate 5,967,328 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 5,967,328 Total Annual Cost (Annualized Capital Cost + Operating Cost) 9,257,704 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 SCR EU042 SCR 28 of 40

102 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 9: -Summary of Utility, Chemical and Supply Costs Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Equivalent Duty 3,462 Plant Cap kw A 355,097 Est power platn eff 35% Unc Nox lb/mmb B 0.11 D=75*(300,000)*((B/1.5)^0.05*(C/100)^0.04)A)^0.35 Watt per Btu/hr Nox Red. Eff. C 80% E=D*A* Equivalent power plant k 355,097 Capital Cost $/kw D $63.11 $22,410, Total SCR Equipment Uncontrolled Nox t/y 1,633.3 Fixed O&M E $147, Annual Operating Hrs 8000 Variable O&M F $598, Uncontrolled Nox lb/mm Ann Cap Factor G 0.82 Heat Input mmb H 6,000 SCR Capital Cost SCR Capital Cost per EPRI Method Duty 3,462 MMBtu/hr Catalyst Area 1,670 ft f (h SCR) Q flue gas 1,603,311 acfm Rx Area 1,921-8 f (h NH 3 ) NOx Cont Eff 80% (as faction) Rx Height 43.8 ft 0 f (h New) new= -728, Retrofit = 0 NOx in 0.11 lb/mmbtu n layer 24 layers Y Bypass? Y or N Ammonia Slip 2 ppm h layer 25.5 ft 127 f (h Bypass) Fuel Sulfur 0.67 wt % (as %) n total 25 layers 29,505,985 f (vol catalyst) Temperature Deg F h SCR 139 ft f (h SCR) Catalyst Volume 122,942 ft 3 New/Retrofit R N or R Electrical Use Duty 3,462 MMBtu/hr kw NOx Cont Eff 80% (as faction) Power 5,133.3 NOx in 0.11 lb/mmbtu n catalyst layers 25 layers Press drop catalyst 1 in H 2 O per layer Press drop duct 3 in H 2 O Total Reagent Use & Other Operating Costs Ammonia Use 56.0 lb/ft 3 Density 146 lb/hr Neat 67.2 gal/hr 29% solution Volume 14 day inventory 22,568 gal $17,062 Inventory Cost 503 lb/hr Design Basis Baseline Em Baseline EmisMax Emis. (Model) Control Eff (%) Cont. Emis (lb/mmbtu) T/yr lb/mmbtu lb/hr 80% 0.02 Nitrous Oxide Actual Method 19 Factor Adjusted Duty 85,334 dscf/mmbtu 9,860 dscf/mmbtu 3,462 MMBtu/hr Operating Cost Calculations Annual hours of operation: 8,688 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.0 hr/8 hr shift 0 0 $/Hr, 0.0 hr/8 hr shift, hr/yr Supervisor 15% of Op. NA - 15% of Operator Costs Maintenance 0.00 Maintenance Total 1.5 % of Total Capital Investment 592,668 % of Total Capital Investment Maint Mtls 0 % of Maintenance Labor NA 0 0% of Maintenance Labor Utilities, Supplies, Rep 0.00 Electricity $/kwh kw-hr 41,474,951 1,658,998 $/kwh, 5,133 kw-hr, hr/yr, 93% utilization Cat. Replacement [14] 35 $/MW-hr mw ,800 Catalyst Replacement 7 Ammonia 0.12 $/lb 503 lb/hr 4,062, ,909 $/lb, 503 lb/hr, hr/yr, 93% utilization See Summary page for notes and assumptions ** Std Air use is 2 scfm/kacfm *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 SCR EU042 SCR 29 of 40

103 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 10: -Summary of Utility, Chemical and Supply Costs Operating Unit: Utility Plant Heater Boiler #2 Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Chemical Engineering Desgin Capacity 400 MMBtu/hr Standardized Flow Rate 615,086 32º F Chemical Plant Cost Index Expected Utiliztion Rate 93% Temperature 121 Deg F 1998/ Expected Annual Hours of Operation 8,688 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 726,000 acfm Inflation Adj 1.19 Expected Equipment Life 20 yrs Standardized Flow Rate 660,092 68º F Dry Std Flow Rate 568,896 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs Purchased Equipment (A) 726,369 Purchased Equipment Total (B) 15% of control device cost (A) 835,324 Installation - Standard Costs 30% of purchased equip cost (B) 250,597 Installation - Site Specific Costs 6,200,000 Installation Total 6,450,597 Total Direct Capital Cost, DC 7,285,921 Total Indirect Capital Costs, IC 31% of purchased equip cost (B) 258,951 Total Capital Investment (TCI) = DC + IC 7,544,872 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 24,884,948 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 1,046,491 Total Annual Cost (Annualized Capital Cost + Operating Cost) 25,931,439 Notes & Assumptions 1 Equipment cost estimate EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 3.2 Chapter Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 3.2 Chapter 2 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat EU042 Reheat 30 of 40

104 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 10: -Summary of Utility, Chemical and Supply Costs CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) 726,369 Purchased Equipment Costs (A) - Absorber + packing + auxillary equipment, EC Instrumentation 10% of control device cost (A) 72,637 ND Sales Taxes 0.0% of control device cost (A) 0 Freight 5% of control device cost (A) 36,318 Purchased Equipment Total (B) 15% 835,324 Installation Foundations & supports 8% of purchased equip cost (B) 66,826 Handling & erection 14% of purchased equip cost (B) 116,945 Electrical 4% of purchased equip cost (B) 33,413 Piping 2% of purchased equip cost (B) 16,706 Insulation 1% of purchased equip cost (B) 8,353 Painting 1% of purchased equip cost (B) 8,353 Installation Subtotal Standard Expenses 30% 250,597 Site Preparation, as required Site Specific 1,400,000 Buildings, as required Site Specific 4,800,000 Site Specific - Other Site Specific 0 Total Site Specific Costs 6,200,000 Installation Total 6,450,597 Total Direct Capital Cost, DC 7,285,921 Indirect Capital Costs Engineering, supervision 10% of purchased equip cost (B) 83,532 Construction & field expenses 5% of purchased equip cost (B) 41,766 Contractor fees 10% of purchased equip cost (B) 83,532 Start-up 2% of purchased equip cost (B) 16,706 Performance test 1% of purchased equip cost (B) 8,353 Model Studies of purchased equip cost (B) 0 Contingencies 3% of purchased equip cost (B) 25,060 Total Indirect Capital Costs, IC 31% of purchased equip cost (B) 258,951 Total Capital Investment (TCI) = DC + IC 7,544,872 Adjusted TCI for Replacement Parts (Catalyst, Filter Bags, etc) for Capital Recovery Cost 7,544,872 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 0.5 hr/8 hr shift, hr/yr 24,603 Supervisor 15% 15% of Operator Costs 3,690 Maintenance Maintenance Labor $/Hr, 0.5 hr/8 hr shift, hr/yr 25,640 Maintenance Materials 100% of maintenance labor costs 256 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 2,690 kw-hr, hr/yr, 93% utilization 869,316 Natural Gas $/mscf, 3,858 scfm, hr/yr, 93% utilization 23,961,443 Total Annual Direct Operating Costs 24,884,948 Indirect Operating Costs Overhead 60% of total labor and material costs 32,514 Administration (2% total capital costs) 2% of total capital costs (TCI) 150,897 Property tax (1% total capital costs) 1% of total capital costs (TCI) 75,449 Insurance (1% total capital costs) 1% of total capital costs (TCI) 75,449 Capital Recovery for a 20- year equipment life and a 7% interest rate 712,183 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 1,046,491 Total Annual Cost (Annualized Capital Cost + Operating Cost) 25,931,439 See Summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat EU042 Reheat 31 of 40

105 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 10: -Summary of Utility, Chemical and Supply Costs Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Catalyst: Catalyst Equipment Life 3 years CRF Rep part cost per unit 0 $/ft 3 Amount Required 39 ft 3 Catalyst Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 Assume Labor = 15% of catalyst cost (basis labor for baghouse replacement) Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit 0 $ each Amount Required 0 Number Total Rep Parts Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 10 min per bag (13 hr total) Labor at $29.65/hr OAQPS list replacement times from 5-20 min per bag. Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Electrical Use Flow acfm P in H2O Efficiency Hp kw Blower, Thermal 726, ,689.8 EPA Cost Cont Manual 6th ed - Oxidizders Chapter Blower, Catalytic 726, ,256.1 EPA Cost Cont Manual 6th ed - Oxidizders Chapter Oxidizer Type thermal (catalytic or thermal) Reagent Use & Other Operating Costs Oxidizers - NA Operating Cost Calculations Annual hours of operation: 8,688 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.5 hr/8 hr shift ,603 $/Hr, 0.5 hr/8 hr shift, hr/yr Supervisor 15% of Op. NA 3,690 15% of Operator Costs Maintenance 0% Maint Labor $/Hr 0.5 hr/8 hr shift ,640 $/Hr, 0.5 hr/8 hr shift, hr/yr Maint Mtls 100% % of Maintenance Labor NA 256 1% of Maintenance Labor Utilities, Supplies, Rep 0% Electricity 0.04 $/kwh kw-hr 21,732, ,316 $/kwh, 2,690 kw-hr, hr/yr, 93% utilization Natural Gas $/mscf 3,858 scfm 1,870,042 23,961,443 $/mscf, 3,858 scfm, hr/yr, 93% utilization See Summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat EU042 Reheat 32 of 40

106 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 10: -Summary of Utility, Chemical and Supply Costs Flue Gas Re-Heat Equipment Cost Estimate Basis Thermal Oxidizer with 70% Heat Recovery Auxiliary Fuel Use Equation 3.19 T wi 121 Deg F - Temperature of waste gas into heat recovery T fi 750 Deg F - Temperature of Flue gas into of heat recovery T ref FER 77 Deg F - Reference temperature for fuel combustion calculations 70% Factional Heat Recovery % Heat recovery section efficiency T wo 561 Deg F - Temperature of waste gas out of heat recovery T fo 310 Deg F - Temperature of flue gas into of heat recovery -h caf Btu/lb Heat of combustion auxiliary fuel (methane) -h wg 0 Btu/lb Heat of combustion waste gas C p wg p wg p af Q wg Btu/lb - Deg F Heat Capacity of waste gas (air) lb/scf - Density of waste gas (air) at 77 Deg F lb/scf - Density of auxiliary fuel (methane) at 77 Deg F 660,092 scfm - Flow of waste gas Q af 3,858 scfm - Flow of auxiliary fuel Year 2005 Inflation Rate 3.0% Cost Calculations 663,950 scfm Flue Gas Cost in 1989 $'s $609,213 Current Cost Using CHE Plant Cost Index $726,369 Heat Rec % A B 0 10, Exponents per equation , Exponents per equation , Exponents per equation , Exponents per equation 3.27 Indurator Flue Gas Heat Capacity - Basis Typical Composition 100 scfm 359 scf/lbmole Gas Composition lb/hr f wt % Cp Gas Cp Flue 28 mw CO 0 v % 0 44 mw CO2 15 v % % mw H2O 10 v % % mw N2 60 v % % mw O2 15 v % % Cp Flue Gas 100 v % % Reference: OAQPS Control Cost Manual 5th Ed Feb Chapter 3 Thermal & Catalytic Incinerators (EPA 453/B ) Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat EU042 Reheat 33 of 40

107 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 11: -Summary of Utility, Chemical and Supply Costs Operating Unit: Line 2 Pellet Induration Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Chemical Engineering Desgin Capacity 400 MMBtu/hr Standardized Flow Rate 615,086 32º F Chemical Plant Cost Index Expected Utiliztion Rate 93% Temperature 121 Deg F 1998/ Expected Annual Hours of Operation 8,688 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 726,000 acfm Inflation Adj 1.19 Expected Equipment Life 20 yrs Standardized Flow Rate 660,092 68º F Dry Std Flow Rate 568,896 68º F CONTROL EQUIPMENT COSTS Capital Costs Year Direct Capital Costs EPRI Correlation 1998 Purchased Equipment (A) ,080,776 Purchased Equipment Total (B) SCR Only 24,581,027 Installation - Standard Costs 15% of purchased equip cost (B) SCR Only 4,083,739 Installation - Site Specific Costs 0 Installation Total 0 Total Direct Capital Cost, DC 0 Total Indirect Capital Costs, IC 0% of purchased equip cost (B) 0 Total Capital Investment (TCI) = DC + IC SCR + Reheat 72,222,298 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. SCR + Reheat 30,618,541 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost SCR + Reheat 7,151,583 Total Annual Cost (Annualized Capital Cost + Operating Cost) SCR + Reheat 37,770,123 Emission Control Cost Calculation Baseline Predicted Limit Cont. Emis. Cont Emis Reduction Cont Cost Pollutant Emis. T/yr lb/hr lb/mmbtu T/yr T/yr $/Ton Rem Nitrous Oxides (NOx) 8, , ,736 Notes & Assumptions 1 Estimated Equipment Cost per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2. 2 For Calculation purposes, duty reflects increased flow rate, not actual duty. 3 Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 4 Capital Cost per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq Reagent Use per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Catalyst Volume per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Reactor Size per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Catalyst Replacement per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Electrical Demand per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq SCR Maintenance Costs EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 4.2 Chapter 2 Eq Control Efficiency = 80% reduction which is the normal SCR control efficiency 12 $35/MW-hr, 140 MW 13 Catalyst replacement every three years. This requires an additional 2 week outage per 3 year outage cycle, annualized to 4.7 days. Notes to User Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 SCR NG EU042 SCR NG 34 of 40

108 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 11: -Summary of Utility, Chemical and Supply Costs CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) Purchased Equipment Costs (A) - Absorber + packing + auxillary equipment, EC 23,080,776 Instrumentation 10% of control device cost (A) NA MN Sales Taxes 6.5% of control device cost (A) 1,500,250 Freight 5% of control device cost (A) NA Purchased Equipment Total (A) 24,581,027 Indirect Installation General Facilities 0% of purchased equip cost (A) 0 Engineering & Home Office 0% of purchased equip cost (A) 0 Process Contingency 0% of purchased equip cost (A) 0 Site Specific-Other 11% Replacement Power, two weeks 2,643,899 Total Indirect Installation Costs (B) 11% of purchased equip cost (A) 2,643,899 Project Contingeny ( C) 15% of (A + B) 4,083,739 Total Plant Cost D A + B + C 31,308,664 Allowance for Funds During Construction (E) Additional 10 week outage for installation 8,232,000 Pre Production Costs (G) 2% of (D+E)) 790,813 Inventory Capital Reagent Vol * $/gal 91,914 Total Capital Investment (TCI) = DC + IC D + E + F + G +H + I 40,423,391 Retrofit multiplier 3 60% of TCI 24,254,035 Adjusted TCI for Replacement Parts (Catalyst, Filter Bags, etc) for Capital Recovery Cost NA Total Retrofit Capital 64,677,426 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator NA - Supervisor NA - Maintenance Maintenance Total 1.50 % of Total Capital Investment 606,351 Maintenance Materials NA % of Maintenance Labor - Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 6,001 kw-hr, hr/yr, 93% utilization 1,939,332 NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - NA NA - Cat. Replacement [14] Catalyst Replacement 548,800 NA NA - Ammonia 0.12 $/lb, 2,708 lb/hr, hr/yr, 93% utilization 2,639,110 NA NA - NA NA - NA NA - Total Annual Direct Operating Costs 5,733,593 Indirect Operating Costs Overhead NA of total labor and material costs NA Administration (2% total capital costs) NA of total capital costs (TCI) NA Property tax (1% total capital costs) NA of total capital costs (TCI) NA Insurance (1% total capital costs) NA of total capital costs (TCI) NA Capital Recovery for a 20- year equipment life and a 7% interest rate 6,105,091 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 6,105,091 Total Annual Cost (Annualized Capital Cost + Operating Cost) 11,838,684 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 SCR NG EU042 SCR NG 35 of 40

109 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 11: -Summary of Utility, Chemical and Supply Costs Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Equivalent Duty 3,462 Plant Cap kw A 355,097 Est power platn eff 35% Unc Nox lb/mmb B 0.58 D=75*(300,000)*((B/1.5)^0.05*(C/100)^0.04)A)^0.35 Watt per Btu/hr Nox Red. Eff. C 80% E=D*A* Equivalent power plant k 355,097 Capital Cost $/kw D $65.00 $23,080, Total SCR Equipment Uncontrolled Nox t/y 8,164.5 Fixed O&M E $152, Annual Operating Hrs 8000 Variable O&M F $626, Uncontrolled Nox lb/mm Ann Cap Factor G 0.82 Heat Input mmb H 6,000 SCR Capital Cost SCR Capital Cost per EPRI Method Duty 3,462 MMBtu/hr Catalyst Area 1,670 ft f (h SCR) Q flue gas 1,603,311 acfm Rx Area 1,921-8 f (h NH 3 ) NOx Cont Eff 80% (as faction) Rx Height 43.8 ft 0 f (h New) new= -728, Retrofit = 0 NOx in 0.58 lb/mmbtu n layer 28 layers Y Bypass? Y or N Ammonia Slip 2 ppm h layer 29.8 ft 127 f (h Bypass) Fuel Sulfur 0.67 wt % (as %) n total 29 layers 34,578,438 f (vol catalyst) Temperature Deg F h SCR 156 ft f (h SCR) Catalyst Volume 144,077 ft 3 New/Retrofit R N or R Electrical Use Duty 3,462 MMBtu/hr kw NOx Cont Eff 80% (as faction) Power 6,000.7 NOx in 0.58 lb/mmbtu n catalyst layers 29 layers Press drop catalyst 1 in H 2 O per layer Press drop duct 3 in H 2 O Total Reagent Use & Other Operating Costs Ammonia Use 56.0 lb/ft 3 Density 785 lb/hr Neat gal/hr 29% solution Volume 14 day inventory 121,571 gal $91,914 Inventory Cost 2708 lb/hr Design Basis Baseline Em Baseline EmisMax Emis. (Model) Control Eff (%) Cont. Emis (lb/mmbtu) T/yr lb/mmbtu lb/hr 80% 0.12 Nitrous Oxide Actual Method 19 Factor Adjusted Duty 85,334 dscf/mmbtu 9,860 dscf/mmbtu 3,462 MMBtu/hr Operating Cost Calculations Annual hours of operation: 8,688 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.0 hr/8 hr shift 0 0 $/Hr, 0.0 hr/8 hr shift, hr/yr Supervisor 15% of Op. NA - 15% of Operator Costs Maintenance 0.00 Maintenance Total 1.5 % of Total Capital Investment 606,351 % of Total Capital Investment Maint Mtls 0 % of Maintenance Labor NA 0 0% of Maintenance Labor Utilities, Supplies, Rep 0.00 Electricity $/kwh kw-hr 48,483,290 1,939,332 $/kwh, 6,001 kw-hr, hr/yr, 93% utilization Cat. Replacement [14] 35 $/MW-hr mw ,800 Catalyst Replacement 7 Ammonia 0.12 $/lb 2708 lb/hr 21,883,299 2,639,110 $/lb, 2,708 lb/hr, hr/yr, 93% utilization See Summary page for notes and assumptions ** Std Air use is 2 scfm/kacfm *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 SCR NG EU042 SCR NG 36 of 40

110 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 12: -Summary of Utility, Chemical and Supply Costs Operating Unit: Utility Plant Heater Boiler #2 Emission Unit Number EU 042 Stack/Vent Number SV 048 & 049 Chemical Engineering Desgin Capacity 400 MMBtu/hr Standardized Flow Rate 615,086 32º F Chemical Plant Cost Index Expected Utiliztion Rate 93% Temperature 121 Deg F 1998/ Expected Annual Hours of Operation 8,688 Hours Moisture Content 13.8% Annual Interest Rate 7.0% Actual Flow Rate 726,000 acfm Inflation Adj 1.19 Expected Equipment Life 20 yrs Standardized Flow Rate 660,092 68º F Dry Std Flow Rate 568,896 68º F CONTROL EQUIPMENT COSTS Capital Costs Direct Capital Costs Purchased Equipment (A) 726,369 Purchased Equipment Total (B) 15% of control device cost (A) 835,324 Installation - Standard Costs 30% of purchased equip cost (B) 250,597 Installation - Site Specific Costs 6,200,000 Installation Total 6,450,597 Total Direct Capital Cost, DC 7,285,921 Total Indirect Capital Costs, IC 31% of purchased equip cost (B) 258,951 Total Capital Investment (TCI) = DC + IC 7,544,872 Operating Costs Total Annual Direct Operating Costs Labor, supervision, materials, replacement parts, utilities, etc. 24,884,948 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 1,046,491 Total Annual Cost (Annualized Capital Cost + Operating Cost) 25,931,439 Notes & Assumptions 1 Equipment cost estimate EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 3.2 Chapter Calculations per EPA Air Pollution Control Cost Manual 6th Ed 2002, Section 3.2 Chapter 2 Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat NG EU042 Reheat NG 37 of 40

111 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 12: -Summary of Utility, Chemical and Supply Costs CAPITAL COSTS Direct Capital Costs Purchased Equipment (A) (1) 726,369 Purchased Equipment Costs (A) - Absorber + packing + auxillary equipment, EC Instrumentation 10% of control device cost (A) 72,637 ND Sales Taxes 0.0% of control device cost (A) 0 Freight 5% of control device cost (A) 36,318 Purchased Equipment Total (B) 15% 835,324 Installation Foundations & supports 8% of purchased equip cost (B) 66,826 Handling & erection 14% of purchased equip cost (B) 116,945 Electrical 4% of purchased equip cost (B) 33,413 Piping 2% of purchased equip cost (B) 16,706 Insulation 1% of purchased equip cost (B) 8,353 Painting 1% of purchased equip cost (B) 8,353 Installation Subtotal Standard Expenses 30% 250,597 Site Preparation, as required Site Specific 1,400,000 Buildings, as required Site Specific 4,800,000 Site Specific - Other Site Specific 0 Total Site Specific Costs 6,200,000 Installation Total 6,450,597 Total Direct Capital Cost, DC 7,285,921 Indirect Capital Costs Engineering, supervision 10% of purchased equip cost (B) 83,532 Construction & field expenses 5% of purchased equip cost (B) 41,766 Contractor fees 10% of purchased equip cost (B) 83,532 Start-up 2% of purchased equip cost (B) 16,706 Performance test 1% of purchased equip cost (B) 8,353 Model Studies of purchased equip cost (B) 0 Contingencies 3% of purchased equip cost (B) 25,060 Total Indirect Capital Costs, IC 31% of purchased equip cost (B) 258,951 Total Capital Investment (TCI) = DC + IC 7,544,872 Adjusted TCI for Replacement Parts (Catalyst, Filter Bags, etc) for Capital Recovery Cost 7,544,872 OPERATING COSTS Direct Annual Operating Costs, DC Operating Labor Operator $/Hr, 0.5 hr/8 hr shift, hr/yr 24,603 Supervisor 15% 15% of Operator Costs 3,690 Maintenance Maintenance Labor $/Hr, 0.5 hr/8 hr shift, hr/yr 25,640 Maintenance Materials 100% of maintenance labor costs 256 Utilities, Supplies, Replacements & Waste Management Electricity 0.04 $/kwh, 2,690 kw-hr, hr/yr, 93% utilization 869,316 Natural Gas $/mscf, 3,858 scfm, hr/yr, 93% utilization 23,961,443 Total Annual Direct Operating Costs 24,884,948 Indirect Operating Costs Overhead 60% of total labor and material costs 32,514 Administration (2% total capital costs) 2% of total capital costs (TCI) 150,897 Property tax (1% total capital costs) 1% of total capital costs (TCI) 75,449 Insurance (1% total capital costs) 1% of total capital costs (TCI) 75,449 Capital Recovery for a 20- year equipment life and a 7% interest rate 712,183 Total Annual Indirect Operating Costs Sum indirect oper costs + capital recovery cost 1,046,491 Total Annual Cost (Annualized Capital Cost + Operating Cost) 25,931,439 See Summary page for notes and assumptions Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat NG EU042 Reheat NG 38 of 40

112 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 12: -Summary of Utility, Chemical and Supply Costs Capital Recovery Factors Primary Installation Interest Rate 7.00% Equipment Life 20 years CRF Replacement Catalyst: Catalyst Equipment Life 3 years CRF Rep part cost per unit 0 $/ft 3 Amount Required 39 ft 3 Catalyst Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 Assume Labor = 15% of catalyst cost (basis labor for baghouse replacement) Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Replacement Parts & Equipment: Equipment Life 3 CRF Rep part cost per unit 0 $ each Amount Required 0 Number Total Rep Parts Cost 0 Cost adjusted for freight & sales tax Installation Labor 0 10 min per bag (13 hr total) Labor at $29.65/hr OAQPS list replacement times from 5-20 min per bag. Total Installed Cost 0 Zero out if no replacement parts needed Annualized Cost 0 Electrical Use Flow acfm P in H2O Efficiency Hp kw Blower, Thermal 726, ,689.8 EPA Cost Cont Manual 6th ed - Oxidizders Chapter Blower, Catalytic 726, ,256.1 EPA Cost Cont Manual 6th ed - Oxidizders Chapter Oxidizer Type thermal (catalytic or thermal) Reagent Use & Other Operating Costs Oxidizers - NA Operating Cost Calculations Annual hours of operation: 8,688 Utilization Rate: 93% Unit Unit of Use Unit of Annual Annual Comments Item Cost $ Measure Rate Measure Use* Cost Operating Labor Op Labor $/Hr 0.5 hr/8 hr shift ,603 $/Hr, 0.5 hr/8 hr shift, hr/yr Supervisor 15% of Op. NA 3,690 15% of Operator Costs Maintenance 0% Maint Labor $/Hr 0.5 hr/8 hr shift ,640 $/Hr, 0.5 hr/8 hr shift, hr/yr Maint Mtls 100% % of Maintenance Labor NA 256 1% of Maintenance Labor Utilities, Supplies, Rep 0% Electricity 0.04 $/kwh kw-hr 21,732, ,316 $/kwh, 2,690 kw-hr, hr/yr, 93% utilization Natural Gas $/mscf 3,858 scfm 1,870,042 23,961,443 $/mscf, 3,858 scfm, hr/yr, 93% utilization See Summary page for notes and assumptions *annual use rate is in same units of measurement as the unit cost factor Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat NG EU042 Reheat NG 39 of 40

113 Cleveland Cliffs Incorporated: United Taconite BART Report - <Appendix A> Emission Control Cost Analysis Table 12: -Summary of Utility, Chemical and Supply Costs Flue Gas Re-Heat Equipment Cost Estimate Basis Thermal Oxidizer with 70% Heat Recovery Auxiliary Fuel Use Equation 3.19 T wi 121 Deg F - Temperature of waste gas into heat recovery T fi 750 Deg F - Temperature of Flue gas into of heat recovery T ref FER 77 Deg F - Reference temperature for fuel combustion calculations 70% Factional Heat Recovery % Heat recovery section efficiency T wo 561 Deg F - Temperature of waste gas out of heat recovery T fo 310 Deg F - Temperature of flue gas into of heat recovery -h caf Btu/lb Heat of combustion auxiliary fuel (methane) -h wg 0 Btu/lb Heat of combustion waste gas C p wg p wg p af Q wg Btu/lb - Deg F Heat Capacity of waste gas (air) lb/scf - Density of waste gas (air) at 77 Deg F lb/scf - Density of auxiliary fuel (methane) at 77 Deg F 660,092 scfm - Flow of waste gas Q af 3,858 scfm - Flow of auxiliary fuel Year 2005 Inflation Rate 3.0% Cost Calculations 663,950 scfm Flue Gas Cost in 1989 $'s $609,213 Current Cost Using CHE Plant Cost Index $726,369 Heat Rec % A B 0 10, Exponents per equation , Exponents per equation , Exponents per equation , Exponents per equation 3.27 Indurator Flue Gas Heat Capacity - Basis Typical Composition 100 scfm 359 scf/lbmole Gas Composition lb/hr f wt % Cp Gas Cp Flue 28 mw CO 0 v % 0 44 mw CO2 15 v % % mw H2O 10 v % % mw N2 60 v % % mw O2 15 v % % Cp Flue Gas 100 v % % Reference: OAQPS Control Cost Manual 5th Ed Feb Chapter 3 Thermal & Catalytic Incinerators (EPA 453/B ) Y:\23\00 MN Taconite BART (2006)\Facility BART Reports\UTAC\United Taconite Control Costs.xls EU042 Reheat NG EU042 Reheat NG 40 of 40

114 Barr Engineering Company Appendix B 4700 West 77th Street Minneapolis, MN Phone: Fax: An EEO Employer Minneapolis, MN Hibbing, MN Duluth, MN Ann Arbor, MI Jefferson City, MO Memorandum To: Margaret McCourtney From: Andrew Skoglund Subject: Revisions per your comments Date: May 16, 2006 Project: Taconite Industry BART Clients c: Mary Jean Fenske, Barr Taconite BART Project Team, Taconite BART Industry Reps. Attached is a revised version of our proposed changes to the BART-analysis modeling protocol. They are set up in the format described in Appendix G to the modeling protocol. The proposed OZONE.DAT files and a figure depicting the proposed modeling domain are also included, as requested. The PSD modeling protocols referenced for the CALMET parameters are based on PSD modeling protocols submitted for Mesabi Nugget LLC, Northshore Mining Line 5 Restart, and Minnesota Steel Industries LLC. Each of these facilities has submitted a modeling protocol using MM4/MM5 data with observations for review. The values noted are representative of those that were used after receiving comment from the FLMs. The Minnesota Steel Industries modeling protocol was submitted in May 2005, with FLM response on June 14, FLMs approved of the submitted values. The comments section regarding receptors has been revised to indicate that we will be using a subset of the original MPCA receptor group, using only BWCA and Voyageurs receptors. Thank you, Andrew J. Skoglund Barr Engineering Co. (952) askoglund@barr.com