Cost Analysis for Compliance with EPA’s Regional NOx Emissions Reductions for Fossil-Fired Power Generation

To achieve a more stringent ambient-air ozone standard promulgated in 1997, the U.S. EPA haa established summer NOx emisaiona limits for fossil-fired electric power generating units in the Ozone Transport Rulemaking region, consisting of 22 eastern and midwestem states and the District of Columbia These jurisdictions are required to submit State Implementation Plans by September 1999 in response to EPA’s rule, with compliance required by 2007. There are 1757 affected units in this region. In the present study, projected state-by-state growth rates for power production are used to estimate power production and NOx emissions by unit in the year 2007. NOx emissions reductions expected by January 1,200O due to Title IV compliance are estimated, leaving a substantial balance of emissions reductions to be achieved by post-combustion NOx control. Cost estimates are developed for achieving these remaining reductions. INTRODUCTION The Clean Air Act Amendments ~,f 1390 (CAAA). administered by the U.S. Environmental Protection Agency (EPA), require reductions in ground-level ozone and its precursors, including nitrogen oxides (NOx). Title IV, being implemented in hvo phases, addresses acidic deposition and establishes point-source NOx omission limits in terms of lb/million Btu of fuel fired. Table 1 gives the NOx emissions limita for Title IV. Title IV standards can generally be met by combustion modifications, whereas Title I limitations will require the use of advanced NOx control technologies. Ozone Regulatory Requirements Title I Title I sets standards for control of six criteria pollutanta, including ozone. Reviewing these standards every five years is mandated. Ground-level ozone is a major ingredient of smog. Since NOx is a major ozone precursor, it is necessary to control NOx to comply with ambient ozone standards. Effective July 16, 1997, the National Ambient Air Quality Standard (NAAQS) for ozone is 0.08 ppm (S-hour average). At this level, many largeand medium-sized urban areas are clasaitied as being in nonattairnnent, and many power plants are situated within these nonattainment areas. Nonattainment of ozone standards reaulta not only from NOx emissions in a given locality but also from significant amounts of NOx transported by winds over a wide geographical area. meet Title IV and SCR and/or SNCR accomplish the final reduction required to meet Title I. Therefore, no further consideration was given to those alternative combustion modification technologies in this study. ECONOMICS METHODOLOGY Budgetary economics were calculated for the selected NOx control strategies, using the ECONMOD computer program developed by FETC [5], which incorporates the methodology for electric power generation costs established by the Electric Power Research Institute (EPRI) [6]. In the present study, costs are reported on a 1997 constant-dollar basis, and are based on the following financial assumptions: 50% debt at 8.5% return, 15% preferred stock at 7.0% return, and 35% common stock at 7.5% return. For a lo-year project life, the corresponding constant-dollar capital charge factor is 0.1557. For a 15-year project life the capital charge factor would be 0.1226, and for a 20-year project life it would be 0.1065. APPROACH OF THIS STUDY The boiler population affected by the SIP Call consists of a total of 1757 fossil-fired power generating units, as identified in a database developed by EPA. A logical sequence of NOx removal processes was assumed in the present study, using SCR and/or SNCR to achieve the required degree of emissions reduction. The less expensive SNCR would be used fust, followed by the more expensive SCR. It was assumed that, at the time the proposed emissions controls are implemented in May 2003, NOx emissions will have been reduced by means of combustion modification to meet Title IV requirements, and that post-combustion control technologies would be used to achieve the f&, more stringent requirements under Title I. For either SNCR or SCR (or their combination), a maximum NH, slip level of 3 ppm was set. Rather than assess each unit individually, a matrix of representative units was defined and evaluated. Costs of installing NOx control technologies were estimated for the representative units. The actual database units were categorized and assigned to thii matrix based on common characteristics, namely plant capacity, fuel type, and NOx level after Title IV controls. To calculate total compliance costs, the representative units were used as proxies for the actual units. Total costs ($/ton) for each unit and technology option were estimated based on age of the power plant, NOx removal percentage, and projected capacity factor. Based on these cost estimates, the least expensive options were progressively implemented until total NOx removal met the seasonal target, of 544,000 tons. Operation of SNCR and/or SCR only during the five-month summertime ozone season was assumed. During the remainder of the year, there would be no consumption of NH,. Capital charges, of course, would continue throughout the year whether or not the SNCWSCR units are operated. Emissions trading was incorporated into the analysis, assuming that any affected electricity generating unit could trade emission allowances with any other affected electricity generating unit without geographic limitations within the SIP Call area. No trading with industrial or other sources was considered. Allowing generating units to over-clean in one ozone season and bank those allowances for use or sale in a foture season was not modeled. For SCR, retrofit difficulty was studied, using retrofit factors of 1.25 and 1.5. Output growth beyond 2007 was not addressed. Since the average ozone-season capacity factor in 2007 for those units predicted to install SCR is SO-90%, it is reasonable to assume that their seasonal heat input, as well as NH3 and catalyst consumption will not change much. Generating more power annually beyond 2007 without increasing total NOx emissions (which are capped) will require more NOx removal from existing units as well as tighter controls on new units. This will impose additional compliance costs not accounted for in this study. On the other hand, retiring existing units and replacing them with new, cleaner units will tend to drive down annual compliance costs. ANALYSIS AND RESULTS Costs at Representative Units Since emissions trading and ideal market conditions were assumed, $/ton of NOx removed was the criterion for selecting units to which NOx control would be applied. The yearly charge for amortized capital (capital charge factor) increases as the life expectancy of the installed equipment decreases, all else being equal. For the base case analysis, a maximum power plant age of 60 years was assumed. Current industry trends show that through one or more refurbishments, many units can be operated for this duration or longer. In 2003, a unit fust placed in service in 1960 would still have 17 years of life before reaching 60 years of age. A control technology would thus be expected to function for 17 years, until the unit’s retirement. This approach to life expectancy is most applicable to SCR installations, which are likely to remain useable even if the boiler is rebuilt as part of a refurbishment. Rebuming and SNCR equipment may be largely lost during a boiler rebuild. Since this analysis shows SCR to be the dominant compliance technology, this method for estimating the life expectancy was not revisited. Assignment of Costs to Database Units SCR on Coal-Fired Units Table 3 shows the SCR capital ($/kW) and levelized O&M; costs ($/ton) assigned to each coal-fired database unit as a fimctionbf Nbx i&t; concentration and unit s&. These C&Y; are based on a 1.50 retrofit factor. For units less than 50 MWe, the 100 MWe representative unit may not be an accurate proxy. Capital costs for these smaller units were extrapolated. The price of anhydrous NH, is $300/tori,, equivalent to $1 lS/ton NOx removed. SCR on Oiland Gas-Fired Units The exhaust gas from oiland gas-fmd boilers and combustion turbines has much less ash and typically less sulfur than that from coal-fired units. This allows the use of higher space velocities (smaller catalyst volumes) and catalysts that are less robust (and presumably somewhat cheaper). The following correlation for SCR on oiland gas-fired boilers and new combined cycle units up to 500 MWe was used [3]: Capital cost ($/kW) = $28.1 * [20O/I~fWj”‘~ At 500 MWe, this correlation predicts a capital cost of $2O/kW, which is also assumed to apply to capacities above 500 MWe. Operating costs include NH3 consumption and catalyst replacement. NH3 price is the same as for coal-fired units. Catalyst replacement costs are assumed to be l/3 those of coal-fued units on a $/ton NOx removed basis, due mostly to smaller catalyst volumes and, to a lesser degree, cheaper catalyst. Costs for oiland gas-fired units are included in Table 3. SNCR on All Units Based on published figures, a capital cost of $15kW was used for all applications of SNCR. Total Regional NOr Removal Costs Base Case 6VO.x Removal: 80%for SCR. 25% for SNCR: SCR Retrofit Factor 1.5) As stated above, the least expensive NOx control options were progressively implemented, unit-byunit, until total NOx emissions met the seasonal target of 543,825 tons. This led to an upper limit or cut-off cost in $/ton. If NOx control at a unit would cost more than this cut-off cost, no controls would be installed at that unit. This analysis projected the cut-off to be about $2,81O/ton. All units for which SCR could be implemented for less than $2,8lO/ton were assumed to apply SCR To these were added SNCR installations at all additional units for which SNCR could be implemented for less than $2,8lO/ton. This method assumes that a power plant owner will remove as much NOx as possible (i.e., choose SCR) if the expected cost ($/ton) is below that of the most expensive NOx control systems being installed. The Base Case results are summarized in Table 4. The average NOx removal cost for all affected units is $l,602/ton. Total NOx removed is 964,643 tonakeason, which is within 0.7% of the 957,975 ton target projected by EPA, and gives aNOx emissions rate of about 537,100 tons/season. Only about 2% of NOx removal is from oilor gas-fired units. Removal costs range from about $7402,800/ton for SCR and about $1,140-2,8OO/ton for SNCR. Average costs are higher for SNCR because most units that would be low-cost SNCR sites are also low-cost SCR sites, and SCR is needed at these units to meet the 64% region-wide reduction target. The average capital charge factor for SCR units is 0.112, vs. 0.139 for SNCR, indicating that SCR is applied to a slightly younger segment of the boiler population (by an average of about 5 years). Hybrids were found to be more expensive ($/ton) than SCR, and so were not chosen for any units. However, at many units, the cost difference was less than 20%, suggesting that more detailed unitspecific analyses may fmd hybrids to be preferred. Case 2 fNOx Removal: 80% for SCR. 40% for SNCR: SCR Retrofit Factor 1.5) The analysis was repeated assuming that SNCR could achieve 40% removal rather than 25%. This lowers the average $/ton cost for SNCR. Some of the NOx removal that was achieved in the Base Case by SCR is achieved by SNCR in Case 2. As shown in Table 5, the average cost drops to $l,526/ton. Marginal costs for additional NOx removal are about $2,65O/ton (vs. about $2,800 in the Base Case). However, over 93% of the NOx removal is still achieved by SCR. Case 3 fN0.x Removal: 80%for SCR. 25% for SNCR: SCR Retrofit Factor 1.25) The analysis was repeated assuming a retrofit factor of 1.25 for SCR installed on coal-fired units, rather than 1.5. No adjustment was made to oilor gas-freed boilers or combustion turbines. Because two-thirds of the annual compliance cost is amortisation of SCR capital costs, and 60-80% of that capital is affected by retrofit complexity, total compliance costs are more sensitive to SCR capital costs than to SNCR performance using the assumptions in this study. As shown in Table 6, the average cost drops to $l,443/ton. Marginal costs for additional NGx removal are about $2,50O/ton. Comparison with EPA’s Cost Estimate In its rulemaking announcement of September 24, 1998, EPA projected a compliance cost of $l,468/ton, which is within the range of $1,443-l,602/ton estimated in the present study. It is difftcult to make direct comparisons between these results since details of the EPA calculations are not provided, such as assumptions regarding capital recovery rate, remaining life of power plants, capital and operating costs for both the SNCR and SCR processes, SCR catalyst life, etc. CONCLUSIONS To achieve the required reduction in NOx emissions in the SIP call area, it is projected that SCRwould be selected for about 500 fossil-tired boilers, totaling about 180 GWe, and SNCR would beselectedfor about 200-300 units, totaling about 20-35 GWe. 98% of the NOx removed is t?om coal-fired units, and 93-98% of all NOx removal is achieved by SCR. The average levelized cost ofcompliance would be $1,400-1,60O/tonof NOx removed (constant 1997 dollar basis), representinga total levelized cost of $1.4-1.6 billion/year. These results are in good agreement with EPA’sprojected costs of $l,468hon of NOx removed and $1.4 billion/year.REFERENCES 1. EPA, Finding of Significant Contribution and Rulemaking for Certain States in the OzoneTransport AssessmentGroup Region for Purposesof Reducing Regional Transport of Ozone,63 FR 90, September 24, 1998. 2. Evaluation of NO, Removal Technologies-Volume 1, Selective Catalytic Reduction, Revision2, Bums and Roe Services Corporation, Submitted to DOEiFederal Energy Technology Center,September 1994. 3. JamesE. Staudt (Andover Technology Partners), Status Report on NOx Control Technologiesand Cost Effectivenessfor Utility Boilers, prepared for Northeast Statesfor Coordinated Air UseManagement(NESCAUM) and Mid-Atlantic Region Au ManagementAssociation (MARAMA),June 1998. 4. Institute of Clean Air Companies (ICAC), Letter to EPA, August 5, 1998. 5. Computer Program for the Economic Evaluation of Emissions Control Technologies, Bums andRoe ServicesCorporation, Submittedto DOE/Federal Energy Technology Center, January 1993. 6. TAG -Technical AssessmentGuide, Electric Power ResearchInstitute, 1993,. Table 1. NOx emissionsregulations for coal-tired boilers under Title IV.