EP A/600/A-96/009 Preliminary Inventory of Hazardous Air Pollutant Emissions from Electric Utility Boilers in the United States Eric S. Ringler Southern Research Institute P.O. Box 13825 Research Triangle Park, NO 27709-3825 Julian W. Jones U.S. Environmental Protection Agency Air Pollution Prevention and Control Division (MD-62) Research Triangle Park, North Carolina 27711 ABSTRACT The Clean Air Act Amendments of 1990 (CAAA) dramatically increased the number and type of sources subject to potential regulation. Title III of the CAAA identifies 189 hazardous air pollutants (HAPs) and indicates that many sources emitting these compounds may ultimately be required to meet strict industry-specific Maximum Achievable Control Technology (MACT) emissions standards. Fossil fuel (i.e., coal, oil, natural gas) combustion generates a variety of organic HAPs and a number of toxic metals (e.g., mercury, chromium, arsenic). Since utility boilers consume a large proportion of the total fossil fuel used in the U.S., including most of the coal, they may be a significant source of HAPs. EPA's Air Pollution Prevention and Control Division (APPCD) has been conducting research to identify and characterize HAP emissions from significant sources, including fossil fuel combustion in electric utility boilers. This paper summarizes an effort undertaken to develop an inventory of HAP emissions from these boilers. Several approaches toward developing and utilizing suitable emission factors are examined, and a HAP inventory is presented that is based on the most current, comprehensive, and quality assured test data available. Analyses are presented that compare results and assess the significance of emissions estimates based on the various approaches examined. INTRODUCTION The Clean Air Act Amendments of 1990 (CAAA) dramatically increased the number and type of sources subject to potential regulation. Title III of the CAAA identifies 189 hazardous air pollutants (HAPs) and indicates that sources liberating significant quantities of these compounds may ultimately be required to meet strict industry-specific Maximum Achievable Control Technology (MACT) emissions standards. Fossil fuel (i.e., coal, oil, natural gas) combustion generates a variety of organic HAPs and a number of toxic metals (e.g., mercury, chromium, and arsenic). Since utility boilers consume a large proportion of the total fossil fuel used in the U.S., including most of the coal, they may be a significant source of HAPs. EPA's Air Pollution Prevention and Control Division (APPCD) has been conducting research to identify and characterize HAP emissions from significant sources, including fossil fuel combustion in electric utility boilers. This paper summarizes an effort undertaken to develop a state-level inventory of HAP emissions from these boilers. The effort was conducted in three phases, each succeeding phase involving more recent, detailed data and information than its predecessor. SUMMARY OF FIRST TWO PHASES The first phase of APPCD's effort was to develop a preliminary inventory based on emission factors adapted from published EPA sources. An activity data base was developed based on the Department of Energy's (DOE's) EIA Form 767 data base for 1991. Because of the large number of boiler firing types and control device types represented in the EIA-767 data, some aggregation of ------- these was necessary to arrive at a reasonable number of model categories for which emission factors were to be developed. This was accomplished by combining boiler and control device types with similar emissions characteristics. The final list of boiler categories is given in Table 1. Then, a literature survey was completed to identify the most complete and appropriate sources of EPA- published emission factors.1"10 These emission factors were then applied to the activity data to develop an initial HAP emissions inventory. A critical review of the emissions inventory resulting from the first phase of the effort suggested that its completeness and accuracy were compromised by the limited availability of representative and reliable emission factors. In addition, experience with measurements of emissions from coal combustion at power plants suggested that some of the emissions estimates generated using the published EPA emission factors were unreasonable. Finally, the effects of FGD systems on HAP emissions had not been addressed, because none of the published EPA emission factors included these effects. The second phase of the effort was a systematic characterization of inorganic HAP emissions due to coal combustion, based on a detailed analysis of the fate of trace elements contained in coal. Trace element data were obtained from the U.S. Geological Survey (USGS) data on coal quality (COALQUAL) in geologic deposits throughout the U.S. Median concentrations of coals in the data base were used as the starting point in estimating HAP emission factors. Coal cleaning, which reduces sulfur and ash from the coal, also reduces the concentrations of toxic elements. The "as-mined" concentrations obtained from the USGS data base were reduced to account for coal cleaning according to the average elemental reductions reported in the literature.11 During combustion, toxic elements are partitioned into three fractions: bottom ash, fly ash, and vapor. Elements with low vapor pressures are generally found in the bottom ash and fly ash, with little or no vapor fraction. Elements with moderate vapor pressures tend to concentrate in the fly ash due fo vaporization in the combustion zone with subsequent condensation as the flue gas is cooled in the heat exchanger. This condensation effect tends to increase the concentration of these elements in the small particle size range; this phenomenon is termed "enrichment." Elements with relatively high vapor pressures can have a significant vapor fraction at exhaust temperatures, with only minor fractions in the fly ash. Elemental partitioning data for pulverized coal and cyclone boilers were obtained from the literature.12 Uncontrolled emission factors can be estimated based on the fuel concentration data and the sum of the fly ash and vapor fractions. Although particulate matter (PM) control devices [e.g., electrostatic precipitators (ESPs)] can reduce the emissions of toxic elements that are emitted as fly ash, these controls have little or no effect on the vapor fraction. Flue gas desulfurization (FGD) can reduce both the fly ash fraction and the vapor fraction of many elements. Information on the fractional removal efficiency of PM control devices was obtained from the literature.13 These removal efficiencies were used to estimate the removal of toxic elements in the fly ash. Controlled emission factors for these devices were then estimated based on the sum of the controlled fly ash and the vapor fraction. In addition, these particulate control devices will have little to no effect on gaseous pollutants such as hydrogen chloride (HC1) and formaldehyde. Two categories of FGD systems exist: wet scrubbers and dry scrubbers. The operational characteristics of these two categories of FGD systems are significantly different. However, insufficient HAP emissions data exist to adequately differentiate between the two technologies." In addition, dry scrubbers are currently used at only 20 of the facilities included in the preliminary inventory, while wet scrubbers are used at 168 facilities. As such, the available HAP removal efficiency data for wet scrubbers were used to estimate emission factors for facilities equipped with either FGD system. By applying the adjustments for the technologies just described, emission factors for 20 inorganic HAPs from coal combustion were calculated based on average fuel concentrations. The average fuel concentrations were partitioned into bottorj/fehi fly ash, HdpGP. Jlifi-fly iJjli WS 2 ------- further partitioned into a series of size fractions between 0 and 30 micrometers. An uncontrolled emission factor was derived based on the sum of the fly ash and vapor fractions. Size-specific control efficiencies were then used to account for HAP removal in common control devices. This approach yielded a more complete set of emission factors (specific to pollutant, boiler type, control type, and fuel) than were attainable in the first phase of this effort (using EPA- published emission factors), and also addressed the effect of FGD systems on HAP emissions. PRELIMINARY INVENTORY DEVELOPMENT USING TEST DATA Emission Factor Development While the first two phases of this study were underway, the Electric Power Research Institute (EPRI) and DOE were conducting field measurements programs to improve available data on power plant emissions of toxic substances. These programs included an extensive array of measurements of airborne emissions at 43 sites selected to be representative of toxic emissions from the U.S. utility industry as a whole. The measurements protocols were carefully researched and documented, consistent and quality assured so that the resulting data represent the most reliable, current, and complete set of direct measurements of HAP emissions from utility boilers available to date. EPRI also made use of USGS data on trace element concentrations occurring in coal to develop relationships between coal concentrations and measurements of emissions of particulate inorganic HAPs.15 The EPRI/DOE data became available after the first two phases of the effort were complete. The third phase adapted the EPRI/DOE test data and emissions relationships to produce a preliminary emissions inventory. The approach was designed to yield emission factors for the model boiler configurations (i.e., boiler type, fuel, and control device) derived in the first phase effort from those present in the 1991 DOE Form EIA-767 activity data. This approach is thus representative of real world plant "configurations and operating conditions, while providing a reasonable summary level of aggregation. It must be emphasized that this inventory and the associated emission factors were developed as part of a research effort to assess alternative approaches for characterizing HAP emissions from the utility industry. The inventory and emission factors are currently under review by EPA regulatory personnel and therefore do not represent EPA policy or guidance. As such, the inventory should be viewed as preliminary. The first step in developing the preliminary inventory was to develop emission factors from the test data using suitable averaging techniques. Although the EPRI/DOE test data represented nearly all of the most significant model boiler configurations, only a few configurations were represented with a sufficient number of tests to provide robust average emission factors. EPRI used two general approaches in an attempt to derive the most representative possible emission factors from the test data. First, for non-volatile inorganic HAP (trace element) emissions from coal -fired boilers, EPRI developed correlations of test values (in pounds per million British thermal units) with plant- specific trace element concentrations and PM emissions. Second, for volatile inorganic HAPs, organic HAPs, and radionuclide emissions from coal-, oil-, and gas-fired boilers, EPRI pooled the available test data into more general categories to provide a more robust data set for determining average emission values. The overall rationale, approach, and statistical methodologies that EPRI used in developing emission factors were reviewed and were found to be satisfactory. Therefore, EPRI's results were adopted as a basis for the preliminary inventory. While EPRI applied the emission factors and correlations on a plant-specific basis, in this study they were applied to model boiler categories. To obtain emission factors for trace element emissions from coal-fired boilers, the EPRI correlations were applied using representative inputs for each model boiler configuration. The average PM input was obtained from PM emissions data reported on EIA Form 767. Average ash values for each configuration were similarly obtained. Trace element concentration inputs specific to 3 ------- coal rank (bituminous, subbituminous, and lignite) were taken as median values for major coal- producing regions in the country based on the USGS COALQUAL data reported by EPRI. The COALQUAL data base represents trace element concentrations in geologic deposits of coal, but does not necessarily represent the quality of coal that is consumed by utility boilers. EPRI screened the COALQUAL data to obtain results that better represent coal that would be supplied to electric utilities. Note also that, in EPRI's inventory, trace element concentration data were used that were specific to the coal supply region likely to supply a given plant with fuel. This level of refinement was not applied in this study because emission estimates were aggregated for model configurations and not specific plants. Emissions were calculated for 19 HAPs, including 9 inorganic HAPs, 4 volatile inorganic HAPs, 5 organic HAPs (including equivalents for total PAH and dioxins/furans), and radionuclides. Specific HAP species are listed below. Inorganic HAPs Antimony (Sb) Arsenic (As) Beryllium (Be) Cadmium (Cd) Chromium (Cr) Cobalt (Co) Lead (Pb) Manganese (Mn) Nickel (Ni) Volatile Inorganic HAPs Mercury (Hg) Selenium (Se) Hydrogen chloride (HC1) Hydrogen fluoride (HF) Organic HAPg Benzene (QH6) Toluene (C6H5CH3) Formaldehyde (HCHO) Polycyclic aromatic hydrocarbons (PAHs) — Benzo(a)pyrene equivalents Dioxins/Furans — 2,3.7,8-TCDD equivalents These HAPs have been identified in power plant stack emissions and have known toxicity factors that can be employed in carrying out a health risk assessment. Preliminary Inventory Estimates of 1991 U.S. Electric Utility Boiler HAP Emissions In 1991, it is estimated that U.S. fossil fueled power plants emitted 1,708 tons* of combined HAPs, excluding HC1 and hydrogen fluoride (HF). Coal-fired boilers made up about 72 percent of these emissions, followed by oil at about 24 percent and gas at 4 percent. HC1 emissions totaled 115,076 tons, and HF emissions totaled 33,373 tons. Coal-fired boilers made up 99 percent of the HC1 and HF emissions, with oil making up the remainder. HAP emissions are summarized in Table 2. These values represent total emissions from all fuels (coal, oil, and gas) consumed by a given boiler, as opposed to emissions associated with the primary fuel only. Many boilers use more than one fuel; e.g., a primary coal-fired boiler may use natural gas during start-up and may use some oil on occasion as makeup fuel. Note that emissions of HC1 and HF overwhelm emissions of all other HAPs combined. This is due to the relative abundance of naturally occurring chlorine and fluorine in coal as well as the relative ineffectiveness of PM control devices in reducing vapor-phase emissions. Of the total HAP emissions for all fuels combined, selenium, nickel, and manganese are emitted in the greatest quantities at 28, 26, and 15 percent of total emissions, respectively (excluding HC1 and HF). Together, these three HAPs make up about 70 percent of the total. The selenium and manganese emissions are almost wholly from coal combustion, while nickel emissions are mostly from oil combustion. Chromium, arsenic, formaldehyde, and lead, which are primarily from coal combustion, are the next most abundant emissions, each with 4 to 6 percent of the total. The bulk of the formaldehyde emissions are due to natural gas combustion, but with significant contributions * 1 ton = 0.907 kg 4 ------- from coal and oil. Emissions of all these seven HAPs make up over 90 percent of the total emissions. Mercury, cobalt, benzene, and toluene are the next most abundant HAPs. Mercury and benzene emissions are primarily associated with coal, cobalt with coal and oil, and toluene with gas, coal, and oil. The remaining HAPs make up less than 1 percent of the total emissions. With the exception of formaldehyde, organics make up a small part of total HAP emissions (less than 5 percent) from coal and oil combustion. Note, however, that toxicity may be a more important consideration than the absolute magnitude of the emissions; e.g., relatively small emissions of dioxins/furans may be of greater concern than emissions of HAPs present in greater abundance. Table 3 gives fuel-specific annual emissions in pounds of HAP emitted per plant. It is clear from the results in the table that, if HC1 and HF emissions are excluded, most plants would likely not be classified as a major source of HAP emissions per Section 112 of the CAAA, based on either the 10-ton limit for any single HAP or the 25-ton limit for combined emissions. Excluding HC1 and HF, total annual HAP emissions from coal-, oil-, and gas-fired plants average about 2.7 tons, slightly greater than about 3.5 tons, and about 0.04 tons, respectively. On the other hand, the Section 112 limits are easily exceeded for HC1 and/or HF emissions at coal- and oil-fired plants. Average per-plant HC1 emissions are on the order of 250 tons per year for coal-fired plants and 12 tons per year for oil-fired plants. Average per-plant HF emissions are about 73 tons per year for coal-fired plants. Factors for HF emissions from oil-fired plants could not be determined based on available data. Similarly, for gas-fired plants, data were unavailable to estimate emission factors for HF and HC1. Note that, if the classification is applied at the boiler rather than the plant level, many boilers would be below the 10-ton limit for HC1 and/or HF. Note also that the emissions in Table 3 are based on the primary fuel consumed by a plant; i.e., emissions associated with combustion of secondary fuels is neglected. This does not seriously impact the representativeness of the per-plant estimates since the vast majority of the heat input at most plants is from the primary fuel. This is especially true since secondary fuels are normally oil or gas, and generally lower HAP emission rates are associated with these fuels. COMPARATIVE ANALYSIS OF RESULTS OF THE THREE PHASES A comparison was made of emission factors developed under: (1) the first phase of the effort, or those previously published by EPA (denoted herein as EPA); (2) the second phase of the effort, or the trace element fate analysis (TEFA); and (3) the third phase of the effort, or the preliminary inventory (PI). For the comparison, only the most common boiler/fuel/control configurations and those which are significant because of their ability to reduce HAP emissions (i.e., those with FGD systems) were included. Note, however, that these configurations made up over 90 percent of the total heat input to the utility industry in 1991. To simplify the analysis of this extensive array of data, a systematic approach was utilized, consisting of (1) broad, order of magnitude comparisons, (2) more restrictive comparisons to determine whether factors agreed within a factor of 2, and (3) analysis of trends toward one method over- or under-predicting another. Finally, total HAP emissions estimates based on the PI approach were compared with estimates based on the published EPA factors. Unfortunately, because of space limitations, all of these emission factors cannot be listed herein. Order of Magnitude Comparisons EPA emission factors for coal-fired boilers generally did not agree well with PI or TEFA emission factors. Less than one-third of the EPA emission factors were within the same order of magnitude as the PI emission factors. Similarly, only about one-third of the EPA emission factors are within the same order of magnitude as the TEFA emission factors. Emission factors for arsenic, beryllium, lead, mercury, and selenium tended to agree better than those for other HAPs in both cases. For oil- fired boilers, the agreement between EPA and PI factors was much better. About 75 5 ------- percent of the factors are within the same order of magnitude, and agreement is relatively consistent across different HAPs. Agreement between PI and TEFA emission factors is generally good. Nearly 90 percent of the factors are within the same order of magnitude, and agreement is consistent among all HAPs. Emission factors for combined fabric filter (FF) and FGD controls generally did not agree as well as for other configurations. Note that this comparison did not include oil- and gas-fired boilers, and also did not include organics since TEFA factors were not developed for these fuels and HAPs. Factor of 2 Comparisons In this more restrictive test, few EPA emission factors compared well with PI or TEFA emission factors. Less than 10 percent of EPA emission factors for coal-fired boilers were within a factor of 2 of either PI or TEFA emission factors. Mercury emission factors showed the best agreement, with some agreement for arsenic, lead, and selenium. For oil-fired boilers, less than 20 percent of EPA factors were within a factor of 2 of the PI emission factors. These included factors for arsenic, chromium, manganese, and nickel. About 40 percent of the PI and TEFA factors agreed within a factor of 2. These represent all of the HAPs where data were available for comparison. This level of agreement was also fairly consistent across fuel/control configurations. The agreement increased to 65 percent when the test was relaxed to a factor of 5. Trends EPA factors were nearly always higher than PI or TEFA factors. Less than 15 percent of PI or TEFA factors are higher than the EPA factors. This may reflect a degree of conservatism in the literature-based values, which are predominantly based on engineering estimates, as opposed to source test data. Since most of the EPA factors differed from the PI and TEFA factors by more than an order of magnitude, this level of conservatism was generally on the order of a factor of 10. Overall, there is no clear trend for PI factors to be higher or lower than TEFA factors. About half of the PI factors were higher, and about half were lower. However, there were notable trends for some fuel/control configurations, and for some HAPs. PI factors for subbituminous coal with ESP control, and lignite coal with ESP, ESP/FGD, and ESP/FF control tended to be less than TEFA factors. In addition, PI factors for cobalt, fluorine, and radionuclides tended to be less than the TEFA factors. PI factors for other configurations and HAPs tend to be neutral or higher than TEFA factors. Given the complexity of the emission factor development processes for these two approaches, it is difficult to determine the exact reasons for these trends. Both approaches make use of trace element concentrations in coal, and the same average trace element concentration values were used in each case. The fundamental difference in the two techniques is the use of emissions relationships based on PM emissions and test data for the PI method versus pure engineering analysis in the TEFA method. Inventory Estimates Table 4 compares inventory estimates for HAP emissions (tons per year, in 1991) based on the PI and EPA factors. The EPA factors yielded a significant overestimate of total U.S. HAP emissions. Total PI emissions (excluding HC1 and HF) are 1708 tons, while total EPA emissions are 14,619 tons. The totals are dominated by HAP emissions due to coal combustion. Thus, total PI emissions for coal are less than 10 percent of the total EPA emissions. However, for oil combustion, PI emissions exceed EPA emissions (407 versus 191 tons, respectively). This is largely due to the bigger PI estimate for nickel. For natural gas combustion, PI emissions were also in excess of EPA emissions; however, this was largely due to the fact that gas emission factors for many HAPs were unavailable in the EPA data. Inventory estimates were not generated based on the TEFA factors; however, total TEFA emissions would be of the same order as the PI estimates based on the general level of agreement in the emission factors for the two methods. 6 ------- UNCERTAINTY OF EMISSION ESTIMATES IN THE PRELIMINARY INVENTORY Significant uncertainty is generally associated with inventory efforts of this kind. Inventory efforts are complex and typically rest on a large number of assumptions, and emissions test data are typically highly variable. Based on a very general assessment of the uncertainty in the PI inventory, it appears that the emission factors and emissions estimates may be accurate to within a factor of about 3 to 5, depending on the specific HAP and control configuration. Uncertainty for HAPs with very small emissions (e.g., PAH and dioxins/furans) is larger due to the greater measurement uncertainties. These estimates are based on variability in the test data combined with consideration of the effects of different approaches to developing and applying emission factors to an inventory. Comparing emission factors developed under the TEFA and PI approaches appears to support these estimates of overall uncertainty in the inventories. These two approaches are methodologically independent; the PI approach being based on test data, and the TEFA approach being based on engineering analysis. The fact that emission factors developed under these two approaches agree within a factor of about 2 to 5 suggests that both approaches are essentially valid and that the uncertainty factor of about a factor of 3 to 5 may not be far off the mark. CONCLUSIONS Previously published EPA emission factors for HAP emissions from electric utility boilers are not generally representative and show poor agreement with analysis of USGS coal quality data and recent field test results. Specifically, only about one-third of the EPA emission factors were of the same order of magnitude as factors developed by the other methods. Agreement was somewhat better (for some boiler/fuel/control device configurations) for arsenic, beryllium, lead, mercury, and selenium. For oil-fired boilers, some agreement was also obtained for nickel, manganese, and chromium. This study provides some useful insights concerning the quality and representativeness of several varying approaches for developing a HAP emissions inventory for utility boilers. The preliminary inventory provides reasonably consistent emission factors and emission estimates, with a reasonable level of uncertainty. The inventory and other results of this effort have been provided by APPCD to EPA's Office of Air Quality Planning and Standards (OAQPS) for possible use in future publications of emission factors and/or other uses they may deem appropriate. REFERENCES 1. Compilation of Air Pollutant Emission Factors, 4th Edition and Supplements, AP-42, U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, 1992. 2. Crosswalk/Air Toxic Emission Factor (XATEF) Database, Version 1.2, U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, 1992. 3. Pope, A.A., PA. Cruse, and C.C. Most, Toxic Air Pollutant Emission Factors - A Compilation for Selected Air Toxic Compounds and Sources, EPA-450/2-88-006a (NTIS PB89-135644), U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, October 1988. 4. SPEC J ATE Database, Version 1.5, U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, 1992. 5. Cruse, P.A., Locating and Estimating Air Emissions from Sources of Benzene, EPA-450/4-84- 007q (NTIS PB88-196175), U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, March 1988. 7 ------- 6. Locating and Estimating Air Emissions from Sources of Chromium, EPA-450/4-84-007g (NTIS PB85-106474), U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, July 1984. 7. Vaught, Locating and Estimating Air Emissions from Sources of Formaldehyde (Revised), EPA-450/4-91-012 (NTIS PB91-181842), U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, March 1991. 8. Locating and Estimating Air Emissions from Sources of Polycyclic Organic Matter (POM), EPA-450/4-84-007p (NTIS PB88-149059), U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, September 1987. 9. . Sink, M.K., Control Technologies for Hazardous Air Pollutants, EPA/625/6-91/014 (NTIS PB92-141373), U.S. Environmental Protection Agency, Center for Environmental Research Information, Cincinnati, OH, June 1991. 10. Brooks, G., Estimating Air Toxics Emissions from Coal and Oil Combustion Sources, EPA- 450/2-89-001 (NTIS PB89-194229), U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, April 1989. 11. Akers, D.J., R.L. Dospoy, C.E. Raleigh, and B. Toole-O'Neil, "The Effects of Coal Cleaning on Predicting Air Toxics Emissions," presented at Electric Power Research Institute-sponsored conference on "The Effects of Coal Quality on Power Plants," Charleston, SC, August 1994. 12. Rizeq, R.G., D.W. Han sell, and W.R. Seeker, "Predictions of Metals Emissions and Partitioning in Coal-Fired Combustion Systems," presented at Electric Power Research Institute-sponsored conference on "Trace Element Transformations in Coal -Fired Power Systems - Trace Elements Workshop," Scottsdale, AZ , April 1993. 13. Debarred, J.L., and R.S. Dahlin, "A Performance Estimation Procedure For Utility Fly Ash Precipitators," EPRI Project 629-5, Electric Power Research Institute, Palo Alto, CA, October 1986. 14. Meij, R., "Trace Element Behavior in Coal-Fired Power Plants," presented at Electric Power Research Institute-sponsored conference on "Trace Element Transformations in Coal -Fired Power Systems — Trace Elements Workshop," Scottsdale, AZ, April 1993. 15. Electric Utility Trace Substances Synthesis Report, EPRI TR-1046-14, Project 3081, Electric Power Research Institute, Palo Alto, CA, November 1994. 8 ------- Table 1. Boiler Configurations Represented in EIA-767 Emission Control Total j Primary i Total Heat Input j: | Fuel Type Boiler PM FGD Plants: Plants! Boilers 10" 12 Btu*i Percent EPRI/DOE Tests ICOAL BIT PC ESP FGD 391 39! 65 1,554! 7.83' 3? iCOAL BIT PC ESP UN 247; 247 j 682 6,995 i 35-24 I 13: iCOAL BIT PC FF FGD 3 I 3! 4 149! 0751 1!' ICOAL BIT PC FF UN 71 7 j 22 160! 0.81 | Ii COAL BIT PC MC FGD 1 1 ; 3 22 j 0.11 j COAL BIT PC VS FGD 3 3: 6 971 0.49 | ' COAL BIT CY ESP FGD 2j 2j 4 88 ] 0.44; 1 i; ICOAL BIT CY ESP UN 17 i 17! 30 388! 1,96! 6 : COAL BIT CY FF UN 1! 1! 2 11 I 0.06! COAL LIG PC ESP FGD 10 j 10; 14 636! 3.20 j 2;: COAL LIG PC ESP UN 12! 12! 23 237! 1.20! 2'! COAL LIG PC FF FGD 2 I 2-; 3 70! 0.35; COAL L!G PC FF UN 1 i 1! 2 19! 0.10: ICOAL LIG PC VS UN 1I 1: 1 26 0.13! iCOAL j LIG jCY ESP FGD 1 ; 1 1 34 i 0.17; „ jCOAL I LIG ICY ESP UN 4 I 4: 4 76 i 0.39; COAL LIG jCY FF FGD 1! 1 i 1 24; 0.12: |COAL iSUB PC ESP !FGD 19' 19j 34 977; 4.92! 1' iCOAL |SUB PC ESP I UN 101 ! 101' 199 2,969' 14.961 2;: iCOAL !SUB PC FF I FGD 121 12' 14 387! 1.95! 3! COAL ISUB PC FF I UN 12: 12! 23 299 i 1.50! 5! ICOAL [SUB PC VS jFGD 7! 7! 16 361 I 1.82; COAL SUB ICY ESP FGD 2: 2: 3 72! 0.36: iCOAL ISUB jCY |ESP jUN 25| 25. 42 382; 1.92! 1; |COAL !SUB CY |MC jUN 1 ! 1 i 1 4- 0.02! iCOAL SUB CY jVS FGD 1; 1' 1 28: 0.14; ! OIL j DIS i ESP i FGD 59! i 95! 7| 0.03 S - j.OIL IDIS ' ESP I UN 258! 661 | 129! 0.65! 3 'OIL I DIS ! FF Ifgd i 11 ! ! 14| 1 : 0.00 j !;OIL IDIS ! FF ! UN ! 10! ! 28 i 1i 0.00! j;OIL idis ; MC iFGD ! 2 I i ej 1 j 0,00 i i OIL i dis ; MC lUN : 21 | ! 36I 139 j 0.70; jiOIL IDIS j UN UN 155 i ! 3251 251 1.26! 15 ;:OIL bis VS ;FGD 9, 14' 1; O.OO: SOIL 1 DIS j VS | UN 11 i 1; 0! 0,00 ||OIL RES ! ESP Ifgd , 1! 2! 0' 0.001 liOIL Ires ; ESP | UN i 151 ! 38| 279; 1.40: SOIL !res ! MC i UN 12! 25; 196! 0.99! SOIL : RES ! UN ; UN 53, ; 91! 147; 0.74! * I-: EGAS ; GAS ; UN !un 404' 10101 2,636 i 13 28 9 . ilFuel Types !iBIT Bituminous Coal !;LIG Lignite Coal iSUB Subbituminous Coal |DIS Distillate Oil jl RES Residual Oil jjBoiler Types j!PC Pulverized Coal Boiler !!CY Cyclone Boiler Emissions Control ESP Electrostatic Precipitator FF Fabric Filter MC Multiclone VS Venturi Scrubber FGD Flue Gas Desulfurization (Wet or Dry Scrubber) UN Uncontrolled PM Particulate Matter *1 Btu = 1.056 kJ 9 ------- Table 2. Summary of 1991 Emissions (tons,* except as noted) Coal Oil Gas All Fuel Antimony 3.3 3.3 Arsenic 90.8 2.71 030 93.8 Beryllium 6.7 0.10 0.0066 6.9 Cadmium 0.7 0.64 0.053 1.4 Chlorine (HC1) 1136953 138037 115075.8 Chromium 102.3 2.56 1.45 106.3 Cobalt 232 18.20 0.11 41.5 Fluorine (HF) 33372.5 33372,5 Lead 68.5 3.44 0.53 72.5 Manganese 247.9 639 0.53 254.8 Mercury 58.7 0.26 0.0011 59.0 Nickel 86.2 354.12 3.16 443.5 Selenium 481.2 1.15 0.013 482.4 - Benzene 30.5 0.63 1.05 32.2 Dioxin/Furan 1.6e-05 4.8e-06 1.6e-06 2.2e-05 Formaldehyde 24.1 11.50 44.81 80.4 PAH 1.4e-02 2.2e-03 1.7e-02 Toluene 11.2 5.69 13.18 30.1 Radionuclides (Ci) 21.2 0.06 21.3 Total (excluding radionuclides) 148303.3 1788.0 65.2 150156.5 Total (tpy) - less HC1/HF 1235.5 407.4 65.2 1708.1 No. of Boilers 1200 1336 1010 2090 No. of Plants 456 523 404 756 * 1 ton = 90? kg 10 ------- Table 3. Average Plant Emissions (pounds* per plant per year, except as noted) Coal Oil Gas Antimony 1429 Arsenic 396.75 45.96 03867 Beryllium 29.47 1.67 0.0084 Cadmium 3.10 10.86 0.0672 Chlorine (HC1) 498653.64 23459.12 Chromium 445.99 43.45 1.8492 Cobalt 101.18 309.19 0.1345 Fluorine (HF) 146109.24 Lead 297.94 58.49 0.6725 Manganese 1076.87 108.63 0.6725 Mercury 257.13 4.50 0.0013 Nickel 376.58 6016.60 4.0347 Selenium 2110.43 19.55 0.0168 - Benzene 133.9 10.8 13000 Dioxin/Furan 7.0e-05 S.le-05 2.0e-06 Formaldehyde 105.7 195.5 57.2000 PAH 6.3e-02 3.7e-02 Toluene 493 96.8 16.8000 Radionuclides (Ci) 4.71e-02 5.76e-04 Total (excluding radionuclides) 650161-5 30381.1 83.1 Total (tpy) - less HCI/HF 5398.7 6922.0 83.1 No. of Primary Fuel Plants 450 102 283 No, of Primary Fuel Boilers 1189 241 660 * 1 pound = 0.454 kg 11 ------- jTable 4. Comparison of PI and EPA Total Emissions (tons' per year) i: Fuel Coal Oil 1 1 Gas f HAP PI EPA | PI: EPA PI i EPA I PI:EPA 1 PI EPA 1PI:EPA j Antimony 31 NA*1 NA na! i NA NA ! [Arsenic 91 392 i 0 23 2.71 1.61 1.65! 0.301 NA ! Beryllium 7 28 i 0,24 0.1 i 0.6 i 0,18 ( 001 NA ;• i Cadmium 1 31 I 0,02 0.6! 7.5! 0.091 0.05 NA [Chlorine (HCI) 113,695 NA 1,380.61 NA I NA NA !• (Chromium 102 3,384 ! 0,031 2.6! 7.1! 0.36 i 1.451 NA j Cobalt 23 NA1 18.21 NA I 0.11 NA 1 Fluorine (HF) 33,373 NA1 NA na! I NA NA ( Lead 69 348 j 0,20 3.4! 9.01 0.381 0.53 NA Manganese 248 5,756 i 0.04 6.4 : 3.4 i 1.86 i 0.53 NA Mercury 59 801 0,74 0 3 i 1.5! 0.181 0.001 NA Nickel 86 2,647 | 0.03 354.1 I 153.3: 2.31 i 3.16 NA I Selenium 481 2141 2.25 1.2: 4,5] 0.26; 0.01 NA 1Benzene 31 NA 0.6! NA f 1.05 NA iDioxWFuran 1.61E-05 NA 4.77E-06 j NA I 1.58E-06 NA Formaldehyde 24 NA 11.5 i NA I 44.81 NA PAH 1.45E-02 135 i 1.07E-04 2.19E-03! 2.38! 9.18E-04 i 0 NA Toluene 11 1,379 i 0.01 5.7 j NA i 13.18 35.14: 0.38 : Total HAP {less HCI/HF) 1,200 12,8791 0,09 390 i 1891 2.07: 7 0! Total HCI/HF 147,068 NA 1,381 ; NA I 0 NA * 1 ton = 907 kg i" NA - Not Available 12 ------- MDMDT dtd T3 ni7o TECHNICAL REPORT DATA JN it lvlrC Jb_ rt X r 1U 1 o (Please read Instructions on the reverse before completing) — 1. REPORT NO. 2. EP A/600/A-96/009 3. REC 4. TITLE AND SUBTITLE Preliminary Inventory of Hazardous Air Pollutant Emissions from Electric Utility Boilers in the United States 5. REPORT DATE 6. PERFORMING ORGANIZATION CODE 7, AUTHOR(S) E. S, Ringler (SoRI) and J. W. Jones 8. PERFORMING ORGANIZATION REPORT NO. 9. PERFORMING ORGANIZATION NAME AND ADDRESS Southern Research Institute P. 0. Box 13825 Research Triangle Park, North Carolina 27709-3825 10. PROGRAM ELEMENT NO. 11. CONTRACT/GRANT NO. 68-02-0062, Tasks 0-011, 1-030, and 2-042 12, SPONSORING AGENCY NAME AND ADDRESS EPA, Office of Research and Development Air Pollution Prevention and Control Division Research Triangle Park, NC 27711 13. TYPE OF REPORT AND PERIOD COVEREO Published paper; 3/93-7/95 14. SPONSORING AGENCY CODE EPA/600/13 15.supplementary NOTES APPCD project officer is Julian W. Jones. Mail Drop 62. 919/ 541-2489. EPA/AWMA Emissions Inventory Conference, Research Triangle Park, NC. 10/11-13/95. 16. abstract-jhe paper summarizes an EPA effort undertaken to develop an inventory of hazardous air pollutant (HAP) emissions from fossil fuel combustion in electric utility boilers. It examines several approaches toward developing and utilizing suit- able emission factors, and presents a HAP inventory that is based on the most cur- rent, comprehensive, and quality assured test data available. It presents analyses that compare results and assess the significance of emissions estimates based on the various approaches examined. The Clean Air Act Amendments (CAAA) dramatically increased the number and type of sources subject to potential regulation. Title III of the CAAA identifies 189 HAPs and indicates that many sources emitting these com- pounds may ultimately be required to meet strict industry-specific Maximum Achie- vable Control Technology (MACT) emissions standards. Fossil fuel (i.e., coal, oil, natural gas) combustion generates a variety of organic HAPs and a number of toxic metals (e.g., mercury, chromium, arsenic). Since utility boilers consume a large proportion of the total fossil fuel used in the U. S., including most of the coal, they may be a significant source of HAPs. 17. KEY WORDS AND DOCUMENT ANALYSIS a. DESCRIPTORS b.IDENTIFIERS/OPEN ENDED TERMS c. COSATI Field/Group Pollution Combustion Inventories Emission Electric Utilities Boilers Fossil Fuels Pollution Control Stationary Sources Hazardous Air Pollu- tants (HAPs) 13B 21B 15E 14G 13 A 21D 18. DISTRIBUTION STATEMENT Release to Public 19. SECURITY CLASS (ThisReport)' Unclassified 21. NO. OF PAGES 20. SECURITY CLASS (This page) Unclassified 22. PRICE EPA Form 2220-1 (9-73) ------- |