# National Emission Standards for Hazardous Air Pollutants for Source Categories; Wool Fiberglass Manufacturing > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/fr%3A97-7214 ## Record - **Collection:** Federal Register - **Document type:** Proposed Rule - **Published:** March 31, 1997 - **Citation:** 62 FR 15228 ## Text SUMMARY: This action proposes national emission standards for hazardous air pollutants (NESHAP) for new and existing sources in wool fiberglass manufacturing facilities. The hazardous air pollutants (HAPs) emitted by the facilities covered by this proposed rule include three metals (arsenic, chromium, lead) and three organic HAPs (formaldehyde, phenol, and methanol). Exposure to these HAPs can cause reversible or irreversible health effects including carcinogenic, respiratory, nervous system, developmental, reproductive, and/or dermal health effects. The EPA estimates the proposed NESHAP would reduce nationwide emissions of HAPs from these facilities by 530 megagrams per year (Mg/ yr) (580 tons per year [ton/yr]), an approximate 30 percent reduction from the current level of emissions. Emissions of particulate matter (PM) would be reduced by an estimated 760 Mg/yr (840 ton/yr) under the proposed NESHAP. The standards are proposed under the authority of section 112(d) of the Clean Air Act (CAA) and are based on the Administrator's determination that wool fiberglass manufacturing facilities may reasonably be anticipated to emit several of the 188 HAPs listed in the draft 112(s) Report to Congress from the various process operations found within the industry. The proposed NESHAP would provide protection to the public by requiring all wool fiberglass plants that are major sources to meet emission standards reflecting the application of the maximum achievable control technology (MACT). DATES: Comments. The EPA will accept comments on the proposed rule until May 30, 1997. Public hearing. Anyone requesting a public hearing must contact the EPA no later than April 21, 1997. If a hearing is held, it will take place at 10 a.m. on April 30, 1997. Persons interested in attending the hearing should call the contact person listed below to verify that a hearing will be held. Request to speak at hearing. Persons wishing to present oral testimony must contact the person listed below (see ADDRESSES) by April 21, 1997. ADDRESSES: Comments. Interested parties may submit written comments (in duplicate, if possible) to Docket No. A-95-24 at the following address: Air and Radiation Docket and Information Center (6102), U.S. Environmental Protection Agency, 401 M Street, SW, Washington, DC 20460. The EPA requests that a separate copy of the comments also be sent to the contact person listed below. Docket. Docket A-95-24, containing supporting information used in developing the proposed standard, is located at the above address in Room M-1500, Waterside Mall (ground floor), and may be inspected from 8:00 a.m. to 5:30 p.m., Monday through Friday. Copies of this information may be obtained by request from the Air Docket by calling (202) 260-7548. A reasonable fee may be charged for copying docket materials. Public hearing. If anyone contacts the EPA requesting a public hearing by the required date (see DATES), the hearing will be held at the EPA Office of Administration Auditorium, Research Triangle Park, North Carolina 27711. Persons interested in presenting testimony should contact Ms. Cathy Coats at (919)541-5422. A verbatim transcript of the hearing and any written statements will be available for public inspection and copying during normal working hours at the EPA's Air and Radiation Docket in Washington, DC. FOR FURTHER INFORMATION CONTACT: For information concerning the proposed regulation, contact Mr. William J. Neuffer, Minerals and Inorganic Chemicals Group, Emission Standards Division (MD-13) U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711, telephone number (919) 541-5435. For information regarding Methods 316 and 318, contact Ms. Rima N. Dishakjian, Emissions, Monitoring, and Analysis Division, telephone number (919) 541-0443. SUPPLEMENTARY INFORMATION: Regulated entities: Entities potentially regulated by this action are those industrial facilities that manufacture wool fiberglass. Regulated categories and entities are shown in Table 1. This table is not intended to be exhaustive, but rather provides a guide for readers regarding entities likely to be regulated by final action on this proposal. This table lists the types of entities that EPA is now aware could potentially be regulated by final action on this proposal. To determine whether your facility is regulated by final action on this proposal, you should carefully examine the applicability criteria in section III.A of this preamble and in Sec. 63.1380 of the proposed rule. If you have any questions regarding the applicability of this action to a particular entity, consult the person listed in the preceding FOR FURTHER INFORMATION CONTACT section. Table 1.--Regulated Categories and Entities ------------------------------------------------------------------------ Entity category Description ------------------------------------------------------------------------ Industrial................................ Wool Fiberglass Manufacturing Plants (SIC 3296). Federal Government: Not Affected ............................ State/Local/Tribal Government: Not ............................ Affected ------------------------------------------------------------------------ The information in this preamble is organized as follows: I. Statutory Authority II. Introduction A. Background B. NESHAP for Source Categories C. Health Effects of Pollutants D. Wool Fiberglass Manufacturing Industry Profile E. Pollution Prevention III. Summary of Proposed Standards A. Applicability B. Emission Limits and Requirements C. Performance Test and Compliance Provisions D. Monitoring Requirements E. Notification, Recordkeeping, and Reporting Requirements IV. Impacts of Proposed Standards A. Applicability B. Air Quality Impacts C. Water Impacts D. Solid Waste Impacts E. Energy Impacts F. Nonair Environmental and Health Impacts G. Cost Impacts H. Economic Impacts V. Selection of Proposed Standards A. Selection of Source Category B. Selection of Emission Sources C. Selection of Pollutants D. Selection of Proposed Standards for Existing and New Sources 1. Background 2. Selection of Floor Technologies 3. Emission Limits E. Selection of Monitoring Requirements F. Selection of Test Methods G. Solicitation of Comments VI. Administrative Requirements A. Docket B. Public Hearing [[Page 15229]] C. Executive Order 12866 D. Enhancing the Intergovernmental Partnership Under Executive Order 12875 E. Unfunded Mandates Reform Act F. Regulatory Flexibility G. Paperwork Reduction Act H. Clean Air Act I. Pollution Prevention Act I. Statutory Authority The statutory authority for this proposal is provided by sections 101, 112, 114, 116, and 301 of the Clean Air Act, as amended (42 U.S.C. 7401, 7412, 7414, 7416, and 7601). II. Introduction A. Background Section 112(c) of the Act directs the Agency to list each category of major and area sources as appropriate emitting one or more of the 189 HAPs listed in section 112(b) of the Act. The EPA published an initial list of source categories on July 16, 1992 (57 FR 31576), and may amend the list at any time. ``Wool Fiberglass Manufacturing'' is one of the 174 categories of sources listed in the notice. As defined in the EPA report, Documentation for Developing the Initial Source Category List (docket item II-A-5), the Wool Fiberglass Manufacturing source category includes any facility engaged in producing wool fiberglass from sand, feldspar, sodium sulfate, anhydrous borax, boric acid, or any other materials. Facilities that manufacture mineral wool from rock, slag, and other similar materials are not included in the source category. On December 3, 1993 (58 FR 63941), EPA published a schedule for the promulgation of standards for the sources selected for regulation under section 112(c) of the Act. According to this schedule, MACT standards for this source category must be promulgated no later than November 15, 1997. In the manufacture of wool fiberglass, molten glass is formed into fibers, which are bonded by an organic resin to produce a wool-like material used primarily for thermal and acoustical insulation. The EPA estimates that at the current level of control, 1,770 Mg/yr (1,950 ton/ yr) of metal HAPs and formaldehyde are emitted from glass-melting furnaces and manufacturing lines in wool fiberglass plants nationwide. The HAPs released from glass-melting furnaces include arsenic, chromium, and lead; an estimated 750 Mg/yr (830 ton/yr) of particulate matter also are emitted. Organic HAPs (formaldehyde, phenol, and methanol) are released from rotary spin (RS) forming, curing, and cooling processes and from flame attenuation (FA) forming and curing processes. B. NESHAP for Source Categories Section 112 of the Act requires that EPA promulgate regulations for the control of HAP emissions from both new and existing major sources. The statute requires the regulations to reflect the maximum degree of reduction in emissions of HAPs that is achievable taking into consideration the cost of achieving the emission reduction, any nonair quality health and environmental impacts, and energy requirements. This level of control is commonly referred to as MACT. For new sources, MACT standards cannot be less stringent than the emission control that is achieved in practice by the best-controlled similar source. [See section 112(d)(3).] The MACT standards for existing sources can be less stringent than standards for new sources, but they cannot be less stringent than the average emission limitation achieved by the best- performing 12 percent of existing sources for categories and subcategories with 30 or more sources, or the best-performing 5 sources for categories or subcategories with fewer than 30 sources. In essence, these MACT standards would ensure that all major sources of air toxic emissions achieve the level of control already being achieved by the better controlled and lower emitting sources in each category. This approach provides assurance to citizens that each major source of toxic air pollution will be required to effectively control its emissions. At the same time, this approach provides a level economic playing field, ensuring that facilities that employ cleaner processes and good emissions controls are not disadvantaged relative to competitors with poorer controls. The control of HAPs is achieved through the promulgation of technology-based emission standards under sections 112(d) and 112(f) and work practice standards under 112(h) for categories of sources that emit HAPs. Emission reductions may be accomplished through the application of measures, processes, methods, systems, or techniques including, but not limited to: (1) Reducing the volume of, or eliminating emissions of, such pollutants through process changes, substitution of materials, or other modifications; (2) enclosing systems or processes to eliminate emissions; (3) collecting, capturing, or treating such pollutants when released from a process, stack, storage or fugitive emissions point; (4) design, equipment, work practice, or operational standards (including requirements for operator training or certification) as provided in subsection (h); or (5) a combination of the above. [See section 112(d)(2).] The EPA may promulgate more stringent regulations to address residual risk that remains after the imposition of controls within 8 years of promulgation of the NESHAP. [See section 112(f)(2).] C. Health Effects of Pollutants The CAA was created, in part, ``to protect and enhance the quality of the Nation's air resources so as to promote the public health and welfare and the productive capacity of its population'' [42 U.S.C. 7401(b)]. This proposed regulation would protect the public health by reducing emissions of HAPs from wool fiberglass manufacturing facilities. This proposed regulation is technology-based, i.e., based on MACT. Emission data collected during development of this proposed NESHAP show that several HAPs are emitted from wool fiberglass manufacturing plants and will be reduced by implementation of the standard. The proposed emission limits would reduce emissions of three particulate metal HAPs: chromium, arsenic, and lead from glass melting furnaces. The organic HAPs (formaldehyde, phenol, and methanol) are emitted from wool fiberglass manufacturing lines and would also be reduced by the proposed standard. In addition to these HAPs and as a result of the control of the metal HAPs, the proposed standard also would reduce emissions of PM, which is regulated under the CAA as a criteria pollutant, and volatile organic compounds (VOC). More information on PM can be found in EPA's criteria document for PM emissions. Following is a summary of the potential health effects caused by exposure to these pollutants. Three metals--arsenic, chromium, and lead--appear on the section 112(b) list of HAPs and are emitted from glass melting furnaces. Long- term inhalation exposure to arsenic is strongly associated with lung cancer, and also irritates the skin and mucous membranes. The EPA has classified arsenic as a Class A, known human carcinogen. The effects of inhaling chromium depend on whether the oxidation state of the metal is trivalent or hexavalent. Trivalent chromium is an essential nutrient, and is substantially less toxic than hexavalent chromium. Both types of chromium irritate the respiratory tract. Hexavalent chromium inhalation is associated with lung cancer, and EPA has classified it as a Class A, known human carcinogen. Data are insufficient to classify trivalent chromium as to human carcinogenicity. [[Page 15230]] Lead exposure damages the central nervous system, especially in children, who may suffer decreased IQ and other neurobehavioral deficits. Children and adults exposed to higher doses of lead may experience anemia, kidney damage, and high blood pressure. The EPA has classified lead as a Class B2, probable human carcinogen, on the basis of reports of kidney tumors in animal studies. (See docket items II-A- 4, II-A-6, II-A-10, II-I-6, II-I-7, II-I-8.) Exposure to formaldehyde, methanol, and phenol irritates the eyes, skin, and mucous membranes and causes conjunctivitis, dermal inflammation, and respiratory symptoms. Formaldehyde exposure has been associated with reproductive effects such as menstrual disorders and pregnancy problems in women workers. The EPA has classified formaldehyde as a Class B1, probable human carcinogen, on the basis of findings of nasal cancer in animal studies, and limited human data. Phenol has been shown to cause damage to the liver, kidney, cardiovascular system, and central nervous system in animal studies. Acute exposure to methanol (usually by ingestion) is well-known to cause blindness and severe metabolic acidosis, sometimes leading to death. Chronic methanol exposure, including inhalation, may cause central disturbances possibly leading to blindness. Data are not sufficient to classify either phenol or methanol as to potential human carcinogenicity. (See docket items II-A-7, II-A-9, II-I-2, II-I-3, II- I-4.) Formaldehyde, phenol, and methanol also are VOCs, which are precursors to ozone formation. Ambient ozone can cause damage to lung tissue, reduction of lung function, and increased sensitivity of the lung to other irritants. Several provisions of the CAA are aimed at reducing emissions of VOC. Additional information on the health effects of ozone are included in EPA's Criteria document, which support the National Ambient Air Quality Standards (NAAQS) for ozone. The EPA does recognize that the degree of adverse health effects can range from mild to severe. The extent and degree to which the health effects may be experienced is dependent upon (1) the ambient concentrations observed in the area (e.g., as influenced by emission rates, meteorological conditions, and terrain), (2) the frequency of and duration of exposures, (3) characteristics of exposed individuals (e.g., genetics, age, pre-existing health conditions, and lifestyles), and (4) pollutant-specific characteristics (e.g., toxicity, half-life in the environment, bioaccumulation, and persistence). D. Wool Fiberglass Manufacturing Industry Profile Wool fiberglass products are primarily used as thermal and acoustical insulation for buildings, automobiles, aircraft, appliances, ductwork, and pipes. Other uses include liquid and air filtration. Approximately 90 percent of the wool fiberglass currently produced is for building insulation products. Wool fiberglass is currently manufactured in the United States by five companies operating 27 plants in 15 states. According to the size definition applied to this industry by the U.S. Small Business Administration (750 company employees or less), none of these firms is classified as a small business. These plants operate a total of 74 manufacturing lines. Wool fiberglass is manufactured in a process that forms thin fibers from molten glass. A typical wool fiberglass manufacturing line consists of the following processes: (1) Preparation of molten glass, (2) formation of fibers into a wool fiberglass mat, (3) curing the binder-coated fiberglass mat, (4) cooling the mat (not always present), and (5) backing, cutting, and packaging. Wool fiberglass manufacturing plants typically contain one or more manufacturing lines. Raw materials for the glass batch are weighed, mixed, and conveyed to the glass melting furnace, which may be gas-fired, electric, or gas and electric combined. The primary component of wool fiberglass is sand, but it also includes varying quantities of feldspar, sodium sulfate, anhydrous borax, boric acid, and many other materials. Cullet, crushed recycled glass, is a primary component in most batches and is required by Executive Order for Federal agency purchases and by law in certain States. Two methods of forming fibers are used in the industry. In the rotary spin (RS) process, centrifugal force causes molten glass to flow through small holes in the wall of a rapidly rotating cylinder. In the flame attenuation (FA) process, molten glass flows by gravity from a small furnace, or pot, to form threads that are then attenuated (stretched to the point of breaking) with air and/or flame. After the fibers are formed, they are sprayed with a binder and collected as a mat on a moving conveyor. The purpose of the binder is to hold the fibers together and its composition varies with product type. Typically, the binder consists of a solution of phenol- formaldehyde resin, water, urea, lignin, silane, and ammonia. The conveyor carries the newly formed mat through an oven for curing of the thermosetting resin and then through a cooling section. Some products do not require curing and/or cooling. FA manufacturing lines do not have cooling processes. No Federal air standards specifically apply to HAP emissions from wool fiberglass production plants. Emission limits for PM in the new source performance standards (NSPS) for glass manufacturing plants (40 CFR part 60, subpart CC) are applicable to gas-fired and modified process glass-melting furnaces in the wool fiberglass industry that were constructed, modified, or reconstructed after June 15, 1979. The NSPS for wool fiberglass insulation manufacturing plants (40 CFR part 60, subpart PPP) limits PM emissions from wool fiberglass insulation manufacturing lines using the RS forming process that were constructed, modified, or reconstructed after February 7, 1984. The NSPS does not require controls for VOC or organic HAPs. As a result of the NSPS and State requirements, PM controls are in place for most glass-melting furnaces. Of the 56 gas and electric furnaces (including gas/electric combinations), 37 are equipped with baghouses or electrostatic precipitators (ESPs). Among those furnaces without add-on controls are 12 electric furnaces that control PM emissions through their design and operation. Controls also are in place for RS manufacturing lines. All 40 RS forming processes control, to varying degrees, organic emissions using one or more of the several process modifications available to this industry. Of the 43 curing ovens, 14 are equipped with a thermal incinerator. Cooling process emissions are uncontrolled for organic HAP emissions. Because of the differences in emissions potential, limitations on the application of process controls, and the dedication of lines to certain product categories, FA forming processes are separated into four subcategories: light density, automotive, heavy density, and pipe products. None of the light density or automotive FA forming processes are equipped with HAP controls. In a few instances, FA forming processes that produce heavy density products, are controlled using process modifications. All FA forming processes producing pipe products use process modifications. None of the 31 curing ovens on FA manufacturing lines are equipped with HAP emission controls. [[Page 15231]] E. Pollution Prevention Pollution prevention is a partial basis for the emission standards for RS and FA manufacturing lines. The emission standard for RS manufacturing lines is formulated as the sum of the MACT floor emission levels for forming, curing, and cooling where process modification is the MACT floor for forming processes, incineration is the MACT floor for curing ovens, and no control is the MACT floor for cooling processes. The emission standards for new and existing FA manufacturing lines producing pipe products and new FA manufacturing lines producing heavy-density products are the sum of the MACT floor emission levels for forming and curing (there are no separate cooling processes on FA manufacturing lines). Process modification is the MACT floor for forming processes and no control is the MACT floor for curing ovens. By formulating the standard as a sum of the individual forming, curing, and cooling MACT floor emission levels for RS manufacturing lines and forming and curing MACT floor emission levels for certain FA manufacturing lines, we have allowed tradeoffs for existing facilities that will accomplish the same environmental results at lower costs and will encourage process modifications and pollution prevention alternatives. According to the industry, new RS manufacturing lines may be able to meet the line standard without the use of costly incinerators with their energy and other environmental impacts, such as increased nitrogen oxides (NOX) and sulfur oxides (SOX) emissions, by incorporating pollution prevention measures. Pollution prevention alternatives will also increase binder utilization efficiency and reduce production costs for industry. In selecting the format of the emission standard for emissions from manufacturing lines, the EPA considered various alternatives such as setting separate emission limits for each process, i.e., forming, curing, and cooling. A line standard gives the industry greater flexibility in complying with the proposed emission limit and is the least costly because industry can avoid the capital and annual operating and maintenance costs associated with the purchase of add-on control equipment. III. Summary of Proposed Standards A. Applicability The proposed NESHAP applies to each of the following existing and newly constructed sources: glass-melting furnaces located at a wool fiberglass manufacturing plant (Standard Industrial Classification [SIC] code 3296), RS manufacturing lines that produce building insulation, and FA manufacturing lines producing pipe insulation. The proposed NESHAP also applies to new FA manufacturing lines producing heavy density products. Facilities that manufacture mineral wool from rock or slag are not subject to the proposed rule but are subject to a separate NESHAP for mineral wool production. Provisions are included in the NESHAP general provisions (40 CFR part 63, subpart A) for the owner or operator to obtain a determination of applicability. A facility that is determined to be an area source would not be subject to the NESHAP. B. Emission Limits and Requirements Emission limits for PM are proposed for glass-melting furnaces. Because the MACT floor for existing and the MACT floor for new glass- melting furnaces are the same, the same emission limit applies to both new and existing sources. Emission limits for formaldehyde also are proposed for each new or existing RS manufacturing line, each new and existing FA manufacturing line producing pipe insulation, and each new FA manufacturing line producing heavy density products. A surrogate approach, where PM serves as a surrogate for HAP metals and formaldehyde serves as a surrogate for organic HAPs, is employed to allow easier and less expensive testing and monitoring requirements. The proposed emission limits are in the same format (mass of emissions per unit of production) as the existing NSPS for glass-melting furnaces and for wool fiberglass plants--kilograms per megagram (kg/Mg) or pound per ton (lb/ton) of glass pulled. Application of the proposed emission limits to the manufacturing line (forming, curing, and cooling) is consistent with the existing NSPS and the use of a kg/Mg (lb/ton) format recognizes that common industry practice is to vent more than one process unit to common ductwork/controls. This format also provides greater flexibility in achieving compliance with the use of pollution prevention measures, especially process modifications that provide the same environmental benefits without the need to purchase add-on control devices. The proposed emission limits are presented in metric units in Table 2(a) and English units in Table 2(b). The proposed emission limits for existing sources are based on the performance of the control technology identified as the MACT floor. The MACT floor for existing glass-melting furnaces is an ESP or a baghouse. Because well-designed and -operated ESPs and baghouses, which are the MACT floor for existing glass-melting furnaces, represent the best technologies available for controlling PM emissions, including HAP metals, the MACT floor for new sources is the same. Table 2(a).--Summary of Proposed Emission Limits for New and Existing Glass-Melting Furnaces and RS and FA Manufacturing Lines in Wool Fiberglass Manufacturing Plants [Metric units] ------------------------------------------------------------------------ Emission limit Process ------------------------------------------- Existing New ------------------------------------------------------------------------ Furnace..................... 0.25 kg of PM per Mg 0.25 kg of PM per Mg of glass pulled. of glass pulled. RS Manufacturing Line....... 0.6 kg of 0.40 kg of formaldehyde per Mg formaldehyde per Mg of glass pulled. of glass pulled. Pipe Insulation Pipe Insulation FA Manufacturing Line....... 3.4 kg of 3.4 kg of formaldehyde per Mg formaldehyde per Mg of glass pulled. of glass pulled. Heavy Density Heavy Density None................ 3.9 kg of formaldehyde per Mg of glass pulled. ------------------------------------------------------------------------ [[Page 15232]] Table 2(b).--Summary of Proposed Emission Limits for New and Existing Glass-Melting Furnaces and RS and FA Manufacturing Lines in Wool Fiberglass Manufacturing Plants [English units] ------------------------------------------------------------------------ Emission limit Process ------------------------------------------- Existing New ------------------------------------------------------------------------ Furnace..................... 0.50 lb of PM per 0.50 lb of PM per ton of glass pulled. ton of glass pulled. RS Manufacturing Line....... 1.2 lb of 0.80 lb of formaldehyde per formaldehyde per ton of glass pulled. ton of glass pulled. Pipe Insulation Pipe Insulation FA Manufacturing Line....... 6.8 lb of 6.8 lb of formaldehyde per formaldehyde per ton of glass pulled. ton of glass pulled. Heavy Density Heavy Density None................ 7.8 lb of formaldehyde per ton of glass pulled. ------------------------------------------------------------------------ The MACT floor for each new or existing RS manufacturing line is represented by the use of process modification(s) for the forming process and a thermal incinerator for each curing oven. The MACT floor for cooling processes on RS manufacturing lines is no control because none of the existing cooling processes are controlled for HAPs. According to the industry, some existing plants will have to upgrade their process modifications on forming in order to meet the proposed emission limit; none will have to install incinerators on curing to comply with the standard. Process modifications are also the basis for the proposed MACT floor for forming processes on each new and existing FA manufacturing line producing pipe insulation and each new FA manufacturing line producing heavy-density products. Because none of the curing processes on FA manufacturing lines are controlled, the MACT floor is no control. C. Performance Test and Compliance Provisions A one-time performance test would demonstrate initial compliance with the proposed emission limits. Under the proposed NESHAP, the owner or operator would measure PM emissions to the atmosphere from affected glass-melting furnaces using EPA Method 5 in 40 CFR part 60, appendix A and Sec. 63.1389 (Test methods and procedures) of the proposed rule. EPA Method 316, ``Sampling and Analysis for Formaldehyde from Stationary Sources in the Mineral Wool and Wool Fiberglass Industries,'' or Method 318, ``Extractive FTIR Method for the Measurement of Emissions from the Mineral Wool and the Wool Fiberglass Industries'' would be used to measure formaldehyde emissions. Methods 316 and 318 are being proposed concurrently with this proposed rule. Using information from the tests, the owner or operator would determine compliance with the applicable emission limit using the instructions and equations in the proposed NESHAP. During the initial performance test, the owner or operator also would monitor and record the glass pull rate of the furnace during each of the three test runs and determine the emission rate for each run in kilograms (pounds) of emission per megagram (ton) of glass pulled (kg/Mg [lb/ton]). A determination of compliance would be based on the average of the three individual test runs. If an ESP is used to control emissions from a glass-melting furnace, the proposed NESHAP requires the owner or operator to establish the ESP operating parameter(s) that will be used to monitor compliance. For example, the secondary voltage of each ESP electrical field may be monitored to determine proper ESP operations. During the initial performance test, the owner or operator would establish the parameters and the range of these parameter values to be used to monitor compliance with the PM emission limit. If a glass-melting furnace is operated without the use of an add-on PM control device, the owner or operator must establish the furnace operating parameter(s) that will be used to monitor compliance. On cold top electric furnaces, for example, the temperature 18 to 24 inches above the glass melt may be used to indicate proper furnace operations. The owner or operator would establish the range of parameter values during the initial performance test to be used to monitor compliance with the PM emission limit. To determine compliance with the proposed emission limits for new and existing RS manufacturing lines, the owner or operator would measure formaldehyde emissions to the atmosphere from forming, curing, and cooling processes and sum the emissions from these processes. For new and existing FA manufacturing lines producing pipe products and for new lines producing heavy-density products, the owner or operator would measure emissions to the atmosphere from the forming and curing processes and sum the emissions. Using information from the tests, the owner or operator would convert the emission test results to the units of the standard using the instructions and equations in the proposed NESHAP. The owner or operator would conduct the initial performance test for each new or existing RS manufacturing line while making building insulation product. Building insulation is defined in the proposed NESHAP as wool fiberglass insulation having a loss on ignition (LOI) of less than 8 percent and a density of less than 0.03 grams per cubic centimeter (g/cm\3\), or 2 pounds per cubic foot (lb/ft\3\). Initial performance tests for FA manufacturing lines would be conducted on new lines while manufacturing heavy-density products (LOI of 11 to 25 percent and a density of 0.01 to 0.05 g/cm\3\ [0.5 to 3 lb/ft\3\]) and on new and existing lines while manufacturing pipe products (LOI of 8 to 14 percent and a density of 0.05 to 0.1 g/cm\3\ [3 to 6 lb/ft\3\]). During performance tests on RS manufacturing lines producing building insulation and certain FA manufacturing lines, the owner or operator would record the LOI of each product for each line tested, the free formaldehyde content of the resin(s) used during the tests, and the binder formulation(s) used during the tests. The performance tests would be conducted using the resin having the highest free formaldehyde content that the owner or operator expects to use on that line. After the performance test, if the owner or operator wants to use a resin with a higher free formaldehyde content or change the binder formulation, another emission test must be performed to demonstrate compliance. If the owner or operator uses forming process modifications to comply, the process parameters (such as binder solids, binder application rate, or LOI) and their associated levels that will [[Page 15233]] be used to monitor compliance must be established during the performance test. After the performance test, if the owner or operator wants to operate the forming process parameters outside the performance test levels, additional performance tests would be required to verify that the source is still in compliance. If a wet scrubbing control device is used to control formaldehyde emissions from an RS manufacturing line producing building insulation or from certain FA manufacturing lines, the owner or operator must establish the operating ranges of the pressure drop across each scrubber, the scrubbing liquid flow rate to each scrubber, and the identity and feed rate of any chemical additive. The owner or operator of a scrubber would also monitor and record the LOI, the free formaldehyde content of the resin used, and the formulation of the binder used during the performance test. If the owner or operator plans to operate the scrubber in such a way that the pressure drop, liquid flow rate, or chemical additive or chemical feed rate exceeds the values established during the performance tests, additional testing must be performed to demonstrate compliance. The proposed rule would allow the owner or operator of RS manufacturing lines and FA manufacturing lines subject to the NESHAP to conduct short-term experimental production runs, where the formaldehyde content or other process parameter deviates from levels established during previous performance tests, without conducting additional performance tests. The owner or operator would have to apply for approval from the Administrator or delegated State agency to conduct such experimental production runs. The application would include information on the nature and duration of the test runs including plans to perform emission testing. Such experimental production runs are important to industry and allow them to develop new products, improve existing products, and determine the effects on product quality and on emissions of process modifications being considered, such as binder reformulation. If a thermal incinerator is used to comply with the proposed emission limit for formaldehyde, the owner or operator would measure the incinerator operating temperature that will be used to monitor compliance. During the initial performance test, the owner or operator would continuously record the incinerator's operating temperature and determine the average temperature during each 1-hour test run. The average of the three test runs would be used to monitor incinerator compliance. D. Monitoring Requirements All owners or operators subject to the proposed NESHAP would submit an operations, maintenance, and monitoring plan as part of their application for a part 70 permit. The plan would include procedures for the proper operation and maintenance of processes and control devices used to comply with the proposed emission limits as well as the corrective actions to be taken when control device or process parameters deviate from allowable levels established during performance testing. The plan would also identify the control device parameters or process parameters to be monitored for compliance, a monitoring schedule, and procedures for keeping records to document compliance. Under the proposed NESHAP, each baghouse used on a glass-melting furnace would have installed a bag leak detection system that is equipped with an audible alarm that automatically sounds when an increase in particulate emissions above a predetermined level is detected. The monitor must be capable of detecting PM emissions at concentrations of 1.0 milligram per actual cubic meter (0.0004 grains per actual cubic foot) and provide an output of relative or absolute PM emissions. Such a device would serve as an indicator of the performance of the baghouse and would provide an indication of when maintenance of the baghouse is needed. An alarm by itself does not indicate noncompliance with the PM emission limit. An alarm would indicate an increase in PM emissions and trigger an inspection of the baghouse to determine the cause of the alarm. The owner or operator would initiate corrective actions according to the procedures in their operations, maintenance, and monitoring plan. The source would be considered out of compliance upon failure to initiate corrective actions within 1 hour of the alarm. If the alarm is activated for more than 5 percent of the total operating time during the 6-month reporting period, the owner or operator must implement a Quality Improvement Plan (QIP) consistent with subpart D of the draft approach to compliance assurance monitoring.1 --------------------------------------------------------------------------- \1\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- For each ESP controlling PM emissions from a glass-melting furnace, the owner or operator would submit as part of their operations, maintenance, and monitoring plan, a description of how the ESP is to be operated and maintained, the ESP parameter(s) to be monitored, a monitoring schedule, and recordkeeping requirements that document compliance. Corrective action would be taken if the range of acceptable values for the selected ESP operating parameter(s), such as secondary voltage, established during the initial performance test is exceeded based on any 3-hour average of the monitored parameter. A deviation outside the established range would trigger an inspection of the control device to determine the cause of the deviation and to initiate corrective actions according to the procedures in the facility's operations, maintenance, and monitoring plan. Failure to initiate corrective actions within 1 hour of the deviation would be considered noncompliance. If the ESP parameter values are outside the range established during the performance test for more than 5 percent of total operating time in a 6-month reporting period, the owner or operator would implement a QIP consistent with subpart D of the draft approach to compliance assurance monitoring.2 If the ESP parameter values are outside the range for more than 10 percent of total operating time in a 6-month reporting period, the owner or operator would be in violation of the standard. --------------------------------------------------------------------------- \2\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- Under the proposed NESHAP, the owner or operator of a glass-melting furnace whose emissions are not exhausted to an air pollution control device for PM control, would submit as part of their operations, maintenance, and monitoring plan a description of how the furnace is to be operated and maintained, the furnace parameter(s) to be monitored for compliance purposes, a monitoring schedule, and recordkeeping requirements that document compliance. Corrective action would be taken if the range of acceptable values for the selected operating parameter(s), such as air temperature above the glass melt in a cold top electric furnace, established during the initial performance test is exceeded based on any 3-hour average of the monitored parameter. A deviation outside the established range would trigger an inspection of the glass-melting furnace to determine the cause of the deviation and to initiate corrective actions according to the procedures in the facility's operations, maintenance, and monitoring plan. Failure to initiate corrective actions within 1 hour of the deviation would be considered noncompliance. If the furnace operating [[Page 15234]] parameter values are outside the range established during the performance test for more than 5 percent of total operating time in a 6-month reporting period, the owner or operator would implement a QIP consistent with subpart D of the draft approach to compliance assurance monitoring.3 If the furnace parameter values are outside the range for more than 10 percent of total operating time in a 6-month reporting period, the owner or operator would be in violation of the standard. --------------------------------------------------------------------------- \3\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- Under the proposed NESHAP, the owner or operator would continuously monitor and record the glass pull rate on all existing and new glass- melting furnaces. The exception to this would be existing furnaces that do not have continuous monitoring equipment. Such furnaces would measure the glass pull rate at least once per day. If the pull rate exceeds by more than 20 percent the average glass pull rate measured during the performance test, the owner or operator must initiate corrective actions within 1 hour. If the glass pull rate exceeds (by more than 20 percent) the average established during the performance test for more than 5 percent of the total operating time in a 6-month reporting period, a QIP must be implemented consistent with subpart D of the draft approach to compliance assurance monitoring. 4 If the glass pull rate exceeds (by more than 20 percent) the average established during the performance test for more than 10 percent of the total operating time in a 6-month reporting period, it is a violation of the standard. Under the proposed NESHAP, the owner or operator would be allowed to do additional performance testing to verify compliance while operating at glass pull rates that exceed the level established during the initial performance test. The additional performance testing would be required to demonstrate compliance with the applicable formaldehyde emission limits for the affected manufacturing line only. --------------------------------------------------------------------------- \4\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- RS manufacturing lines that produce building insulation and certain FA manufacturing lines would monitor and record the free formaldehyde content of each resin lot, the binder formulation of each batch, and product LOI at least once each day. If resin-free formaldehyde content exceeds the performance test levels, the owner or operator would be in violation of the standard. Under the proposed NESHAP, the binder formulation must not deviate from the formulation specifications used during the performance test. An owner or operator of affected RS or FA manufacturing lines that use process modifications to comply with the emission standard would include in their written operations, maintenance, and monitoring plan how the process will be operated and maintained and identify the process parameters to be monitored, a monitoring schedule, and recordkeeping requirements that document compliance. Examples of process parameters that might be used to monitor compliance include product LOI, binder solids, and binder application rate. The plan would also have to demonstrate that the parameter(s) to be monitored correlate with formaldehyde emissions. The plan would include procedures for establishing maximum or minimum values, as appropriate, based on initial performance testing. Should the process parameter(s) deviate from the range established during the performance test, the owner or operator must inspect the process to determine the cause of the deviation and initiate corrective action within 1 hour of the deviation. If the process parameter(s) is outside the performance test range for more than 5 percent of total operating time during a 6-month reporting period, the owner or operator would implement a QIP consistent with subpart D of the draft approach to compliance assurance monitoring. 5 If the process parameter(s) is outside the range for more than 10 percent of total operating time in a 6-month reporting period, the owner or operator would be in violation of the standard. --------------------------------------------------------------------------- \5\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- An owner or operator who uses a wet scrubbing control device to control formaldehyde emissions from an RS manufacturing line producing building insulation or from certain FA manufacturing lines would continuously monitor and record the pressure drop across each scrubber, the scrubbing liquid flow rate to each scrubber, and the identity and feed rate of any chemical added to the scrubbing liquid. Under the proposed monitoring provisions, corrective action would be taken if any 3-hour average scrubber parameter is outside the range of acceptable values established during the initial performance test. If there was a deviation outside the established range, the owner or operator would inspect the process to determine the cause of the deviation and to initiate corrective actions according to the procedures in the facility's operations, maintenance, and monitoring plan. The owner or operator of the scrubber would be out of compliance upon failure to initiate corrective actions within 1 hour of the deviation. If any scrubber parameter is outside the performance test range for more than 5 percent of the total operating time in a 6-month reporting period, the owner or operator would implement a QIP consistent with subpart D of the draft approach to compliance assurance monitoring. 6 If any scrubber parameter is outside the range for more than 10 percent of total operating time in a 6-month reporting period, the owner or operator would be in violation of the standard. --------------------------------------------------------------------------- \6\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- If an incinerator is used to control formaldehyde emissions from a manufacturing line or from individual forming or curing processes, the owner or operator would continuously monitor and record the operating temperature of each incinerator. The temperature monitoring device would be installed in the incinerator firebox. This is typically done using a thermocouple (a standard feature on most incinerators) and a strip chart recorder or data logger. Following the initial performance test, the owner or operator must maintain the temperature so that the temperature, averaged over a 3-hour period, does not fall below the average temperature established during the initial performance test. A temperature below the performance test average would be considered a violation of the standard. The owner or operator may modify any of the control device or process parameter levels established during the initial performance tests for compliance monitoring. The proposed NESHAP contains provisions that would allow the owner or operator to change add-on control device and process parameter values from those established during the initial performance tests by performing additional emission testing to verify compliance. As required by the NESHAP general provisions (40 CFR part 63, subpart A), the owner or operator must develop and implement a separate startup, shutdown, and malfunction plan. The plan would include procedures for the inspection and determination of the cause of a process or control device malfunction and the corrective procedures to be followed to remedy the malfunction. E. Notification, Recordkeeping, and Reporting Requirements All notification, recordkeeping, and reporting requirements in the general [[Page 15235]] provisions would apply to wool fiberglass manufacturing facilities. These include: (1) initial notification(s) of applicability, notification of performance test, and notification of compliance status; (2) a report of performance test results; (3) a startup, shutdown, and malfunction plan with semiannual reports of any reportable events; and (4) semiannual reports of deviations from established parameters. If deviations from established parameters are reported, the owner or operator must report quarterly until a request to return the reporting frequency to semiannual is approved. In addition to the requirements of the general provisions, the owner or operator would maintain records of the following, as applicable: (1) Bag leak detection system alarms, including the date and time, with a brief explanation of the cause of the alarm and the corrective action taken; (2) ESP monitoring plan parameter values, such as the secondary voltage of each electrical field, for each ESP used to control PM emissions from a glass-melting furnace, including any period when the parameter values deviate from those established during the performance test, with a brief explanation of the cause of the deviation and the corrective action taken; (3) Uncontrolled glass-melting furnace operating parameter values, such as the temperature readings taken above the molten glass in cold top electric furnaces, including any period when the operating parameter values deviate from those established during the performance test, with a brief explanation of the cause of the deviation and the corrective action taken; (4) The LOI and product density for each bonded product manufactured on an RS or FA manufacturing line subject to this NESHAP; (5) The free formaldehyde content of each resin lot and the binder formulation of each batch used in the production of bonded wool fiberglass on RS or FA manufacturing lines subject to this NESHAP; (6) Process parameters for RS and FA manufacturing lines that comply with the emission standards using process modifications, including any period when the parameter levels deviate from levels established during the performance test and the corrective actions taken; (7) Scrubber pressure drop, scrubbing liquid flow rate, and any chemical additive (including chemical feed rate to the scrubber), including any period when the parameter levels deviate from those established during the performance tests and the corrective action taken, (8) Incinerator operating temperature, including any period when the temperature falls below the average level established during the performance test, with a brief explanation of the cause of the deviation and the corrective action taken; (9) Glass pull rate including any period when the pull rate exceeded the average pull rate established during the performance test by more than 20 percent with a brief explanation of the cause of the exceedance and the corrective action taken. Initial performance tests and compliance assurance monitoring requirements for forming process modifications apply only when building insulation products are being manufactured on RS manufacturing lines and when pipe products are being manufactured on new and existing FA manufacturing lines and heavy-density products are being manufactured on new FA manufacturing lines. The LOI must be monitored to demonstrate to EPA the products being manufactured and which lines are subject to the standard. During periods when other products are being manufactured, it is expected that the parameter values, such as LOI or binder solids, may vary from those levels established during the initial performance tests for building insulation on RS manufacturing lines and heavy-density or pipe products on FA manufacturing lines. The NESHAP general provisions (40 CFR part 63, subpart A) require that records be maintained for at least 5 years from the date of each record. The owner or operator must retain the records onsite for at least 2 years but may retain the records offsite the remaining 3 years. The files may be retained on microfilm, on microfiche, on a computer, on computer disks, or on magnetic tape disks. Reports may be made on paper or on a labeled computer disk using commonly available and compatible computer software. IV. Impacts of Proposed Standards A. Applicability All plants in the industry would be subject to the proposed NESHAP unless the owner or operator demonstrates the facility is not a major source according to the requirements in the NESHAP general provisions. Seven of the 30 electric or gas/electric combination glass-melting furnaces are not controlled and are expected to need to install a baghouse or ESP to comply with the proposed emission limit. All gas- fired glass-melting furnaces are well controlled and are expected to be in compliance with the NESHAP. Certain uncontrolled glass-melting furnaces, such as cold top electric furnaces, maintain low PM emissions as a result of their design and operation and are expected to meet the emission limits without the addition of control devices. Some RS forming processes would need to upgrade their process modifications to meet the emission limits for manufacturing lines. B. Air Quality Impacts (Docket Item II-B-22) Most of the existing glass-melting furnaces are already well controlled. At the current high level of control, nationwide emissions of PM are about 750 Mg/yr (830 ton/yr). Because of the existence of controls on all gas furnaces and the emission limiting design and operation of cold top electric furnaces, no emission reduction is expected from gas or cold top electric furnaces under the proposed NESHAP. There are 30 electric or combination gas/electric furnaces of which 23 are well controlled. Under the proposed NESHAP, it is expected that baghouses would be added to the seven uncontrolled electric glass- melting furnaces, which would result in a reduction in nationwide PM emissions of 600 Mg/yr (660 ton/yr) of which 40 Mg/yr (50 ton/yr) is particulate matter less than 10 microns (m) in diameter (PM- 10) (docket item II-B-20). Impacts on new furnaces will vary. New gas- fired glass-melting furnaces would be adequately controlled, even in the absence of the proposed NESHAP, as a result of the NSPS for glass manufacturing plants (40 CFR part 60, subpart CC). Because of their design and operation, new cold top electric furnaces would meet the proposed emission limit for new furnaces without add-on controls. Only new electric furnaces are expected to be impacted by the proposed emission limits for new glass melting furnaces. New electric glass- melting furnaces are not subject to the NSPS for glass manufacturing plants and are likely, under the proposed NESHAP, to need controls to comply with the emission limit for new furnaces. The PM emission reduction from new electric glass-melting furnaces resulting from the proposed emission limit for new furnaces would be 160 Mg/yr (180 ton/ yr) in the fifth year of the standard. Current nationwide emissions of metal HAPs from existing furnaces is 270 kg/yr (600 lb/yr). Under the proposed NESHAP, metal HAP emissions from existing furnaces and new furnaces would be reduced by 9 kg/ [[Page 15236]] yr (20 lb/yr) and 2 kg/yr (5 lb/yr), respectively. Nationwide emissions of formaldehyde from existing manufacturing lines are estimated to be 1,770 Mg/yr (1,950 ton/yr) at the current level of control. Emissions from RS manufacturing lines account for about 70 percent of the formaldehyde emissions. Implementation of the proposed NESHAP would reduce nationwide formaldehyde emissions from existing sources by 410 Mg/yr (450 ton/yr). Emission reductions from RS manufacturing lines producing building insulation constitute the entire reduction; there would be no emission reductions from FA manufacturing lines because, under the proposed emission limits, no additional control of FA manufacturing lines is necessary and no new FA manufacturing lines are anticipated. Reduction in formaldehyde emissions from new RS manufacturing lines is estimated to be 120 Mg/yr (130 ton/yr) in the fifth year of the standard. Nationwide baseline emissions and emission reduction estimates for glass-melting furnaces and manufacturing lines are summarized in metric units in Table 3(a) and in English units in Table 3(b). Table 3(a).--Nationwide Annual Emissions [Metric units] ---------------------------------------------------------------------------------------------------------------- Baseline Emission Source Pollutant emissions reduction (Mg/yr) (Mg/yr)a ---------------------------------------------------------------------------------------------------------------- Glass-Melting Furnaces....................... Metal HAP............................ 0.3 0.01 PM................................... 750 760 RS Manufacturing Lines....................... Formaldehyde......................... 1,220 530 FA Manufacturing Lines....................... Formaldehyde......................... 550 0 All Sources.................................. Total HAPs........................... 1,770 530 PM (Non-HAP)......................... 750 760 Total Pollutants..................... 2,520 1,290 ---------------------------------------------------------------------------------------------------------------- a Emission reduction in the fifth year of the standard. Includes emission reductions from new sources. Table 3(b).--Nationwide Annual Emissions [English units] ---------------------------------------------------------------------------------------------------------------- Baseline Emission Source Pollutant emissions reduction (ton/yr) (ton/yr)a ---------------------------------------------------------------------------------------------------------------- Glass-Melting Furnaces....................... Metal HAP............................ 0.3 0.01 PM................................... 830 840 RS Manufacturing Lines....................... Formaldehyde......................... 1,350 580 FA Manufacturing Lines....................... Formaldehyde......................... 600 0 All Sources.................................. Total HAPs........................... 1,950 580 PM (Non-HAP)......................... 830 840 Total Pollutants..................... 2,780 1,420 ---------------------------------------------------------------------------------------------------------------- a Emission reduction in the fifth year of the standard. Includes emission reductions from new sources. An analysis of emissions from a medium-sized (27,200 Mg/yr [30,000 ton/yr] capacity) model electric furnace shows that metal HAP emissions would be reduced by about 0.001 Mg/yr (0.001 ton/yr) and PM emissions by an estimated 67 Mg/yr (74 ton/yr) from both an existing and a new electric furnace over an uncontrolled electric furnace. For a medium model plant (99,800 Mg/yr [110,000 ton/yr] capacity), metal HAP emissions from existing and new electric furnaces would be reduced by 0.004 Mg/yr (0.004 ton/yr) over a plant with uncontrolled electric furnaces; PM emissions would be reduced by an estimated 250 Mg/yr (270 ton/yr). Under the proposed NESHAP, there would be no emission reductions associated with existing gas-fired or cold top electric furnaces because all gas furnaces are already well controlled and no additional controls would be required for cold top electric furnaces to meet the proposed emission limits. Because new gas furnaces would be controlled as a result of the NSPS for glass manufacturing sources (40 CFR part 60, subpart CC), no additional emission reductions from new gas furnaces would occur under the proposed NESHAP. As with existing cold top electric furnaces, new cold top electric furnaces would be able to meet the proposed emission limit without additional control. Based on model line and plant analyses, formaldehyde emissions from a medium-sized (27,200 Mg/yr [30,000 ton/yr] capacity) RS manufacturing line producing building insulation would be reduced by an estimated 8 Mg/yr (9 ton/yr). Emissions of formaldehyde from a medium-sized plant (99,800 Mg/yr [110,000 ton/yr] capacity) containing two large RS manufacturing lines would be reduced by an estimated 30 Mg/yr (33 ton/ yr). Formaldehyde emissions from a new RS manufacturing line would be reduced an estimated 33 Mg/yr (37 ton/yr). No emission reduction would be achieved for new or existing medium-sized FA manufacturing lines producing pipe insulation since there would be no additional controls under the proposed NESHAP. The formaldehyde emission reduction from a new medium-sized (1,800 Mg/yr [2,000 ton/yr] production capacity) FA manufacturing line producing heavy-density products would total about 2.8 Mg/yr (3.1 ton/yr) although no new FA manufacturing lines are projected. Additional information on model plants and lines is included in the docket. Because EPA proposes to regulate formaldehyde emissions as a surrogate measure for organic HAP emissions from manufacturing lines, only formaldehyde emissions data are presented here, although when the formaldehyde emission limit is met, phenol and methanol emissions will also be reduced. Where incineration is used to control formaldehyde emissions [[Page 15237]] from curing, emissions of phenol and methanol will also be controlled. Emissions data to quantify the degree of reduction in emissions of phenol and methanol as a result of increased levels of forming process modifications are not available. The results of emissions tests conducted at wool fiberglass manufacturing plants, including phenol and methanol test results, are contained in the docket. C. Water Impacts Because this standard is based on the use of baghouses, dry ESPs, thermal incinerators, and process modifications, there are no water pollution impacts. A few existing emission sources may use scrubbers to control HAP emissions although no additional sources are expected to add wet scrubbers for the control of HAP emissions. Therefore, no water impacts are expected from the proposed rule. D. Solid Waste Impacts The PM captured by the baghouses added to the seven uncontrolled electric furnaces will be recycled to the furnace and no solid or hazardous waste is generated by the use of thermal incinerators. No solid waste impacts are expected from the proposed rule. E. Energy Impacts (Docket Item II-B-22) Baghouses require electrical energy to operate fans. The additional electrical energy requirements are estimated to be 1.8 thousand megawatt hours per year (MWh/yr) over current requirements for seven additional baghouses to be added to existing sources. Emissions of PM associated with the additional energy requirements are estimated to be 0.1 ton/yr as compared to the PM emission reduction of 700 ton/yr estimated for installing the seven baghouses on uncontrolled furnaces. Projected new RS manufacturing lines would comply with the proposed standard for new sources using process modifications on forming and incinerators on curing. An additional 2.9 thousand MWh/yr for electricity and 290 billion Btu/yr of natural gas would be required for new incinerators although process modifications only may be used to comply with the proposed standard for new RS manufacturing lines. The total additional energy required as a result of this proposed NESHAP is 300 billion Btu/yr in the fifth year of the standard. No new FA manufacturing lines are projected; thus there are no increased energy requirements under the proposed standard for new FA manufacturing lines. F. Nonair Environmental and Health Impacts Reducing HAP levels may help lower occupational exposure levels and site-specific levels of PM and VOCs. New or upgraded process modifications for forming operations would decrease the quantity of HAP constituents in binder formulations. The addition of baghouses, ESPs, and incinerators may increase noise levels in the plant area due to the operation of pollution control devices where none are currently in place. G. Cost Impacts The EPA analyzed the cost impacts of the proposed standards for glass-melting furnaces by developing model lines based on site-specific information included in the ICR survey responses (docket item II-B-21) coupled with cost algorithms from the OAQPS Cost Manual (docket item II-A-3). The cost impacts of the proposed standards on wool fiberglass manufacturing facilities are based on estimates supplied by wool fiberglass companies for each of their manufacturing lines (docket item II-D-65). The total nationwide capital and annual costs for existing glass- melting furnaces under the proposed NESHAP are $3.2 million and $1.5 million, respectively. This represents the cost of adding baghouses to seven electric glass-melting furnaces as well as the monitoring costs of bag leak detection systems installed on baghouses and temperature monitors installed on cold top electric furnaces. Control cost estimates assume the addition of pulse jet baghouses with polyester filter bags, an air-to-cloth ratio of 0.9 actual cubic meters per minute per square meter (3 acfm/ft\2\), and a pressure drop of 20 cm (8 in.) of water column. The estimated capital and annual costs of control equipment for a medium electric furnace (production capacity of 30,000 ton/yr) are $432,000 and $209,000, respectively. The capital cost includes the cost of the control device, auxiliary equipment, and installation, and retrofit costs. The model furnace cost estimates do not include the capital and annual costs for a bag leak detection system required on all baghouses under the proposed NESHAP. The EPA estimates the capital cost of this monitoring system to be approximately $9,100 per furnace, with $1,800/yr in annual costs. Cold top electric furnaces would incur costs for monitoring an operating parameter that gives an indication of furnace performance; for cost estimating purposes, the cost of monitoring the air temperature above the molten glass surface was used. The estimated capital and annual costs of monitoring the temperature of cold top electric furnaces are $1,500 and $240, respectively. For ESPs, owners or operators are expected to monitor ESP parameters that they commonly monitor, such as secondary voltage, so that no additional monitoring costs would be incurred. Because the NSPS for glass manufacturing sources would regulate any new gas furnaces, there would be no additional control costs for new gas furnaces under the proposed NESHAP. The NSPS for glass manufacturing sources does not cover electric furnaces. Thus, under the proposed NESHAP, new electric furnaces will incur the cost associated with adding baghouses as well as bag leak detection monitoring systems. The capital and annual costs associated with a new baghouse would be $288,000 and $189,000, respectively in addition to the capital and annual costs of a bag leak detection system, $9,100 and $1,800, respectively. Based on information supplied by the North American Insulation Manufacturers Association (NAIMA), 30 RS forming operations would upgrade their proprietary process modifications to meet the proposed emission limit for RS manufacturing lines; none of the existing curing ovens that are uncontrolled for HAPs would have to add an incinerator. No control costs are associated with complying with the proposed NESHAP for FA manufacturing lines. The proposed monitoring requirements for RS and FA manufacturing lines, i.e., monitoring resin free-formaldehyde content, product LOI and density, other process parameters, and incinerator operating temperature, are current industry practices and would not impose any additional costs. However, NAIMA estimates that there would be a one-time cost per line for testing that would be needed to establish a correlation between formaldehyde emissions and the process parameters to be monitored. NAIMA estimated the costs of complying with the proposed standard for RS manufacturing lines for each of their lines. Capital costs per line ranged from $150,000 to $4 million and annual expenses per line ranged from $100,000 to $400,000. Nationwide capital costs of upgrading process modifications on 30 RS manufacturing lines were estimated at $16.3 million with annual costs of $4.8 million. Annual cost for new RS manufacturing lines is estimated to be $0.9 million per line. No FA lines would require additional controls under the proposed standard and there would be no additional control costs. For all RS and FA manufacturing lines subject [[Page 15238]] to the standard, there would be a one-time cost of $15,000 per line to establish the process parameter values for compliance monitoring. Because the process parameters that are likely to be used for compliance monitoring are ones that industry currently monitors, no additional costs will be incurred for monitoring beyond the one-time cost of $15,000 per line. Total nationwide capital costs for the standard are estimated at $19.5 million and annual nationwide costs are estimated at $6.3 million/yr, including installation, operation, and maintenance of emission control and monitoring systems. H. Economic Impacts (Docket Item II-A-12) The economic analysis of the proposed NESHAP finds impacts at the facility and market-level to be modest. The average market price increases for both structural and nonstructural wool fiberglass would be less than 0.5 percent. The resultant decreases in quantity demanded range from 0.17 percent for structural insulation markets to 0.22 percent for nonstructural insulation markets. None of the affected firms are classified as small businesses and no closures are predicted. For more detail, see the full economic impact analysis in the docket. V. Selection of Proposed Standards A. Selection of Source Category Section 112(c) of the Act directs the Agency to list each category of major and area sources, as appropriate, emitting one or more of the 189 HAPs listed in section 112(b) of the Act. The EPA published an initial list of source categories on July 16, 1992 (57 FR 31576), and may amend the list at any time. ``Wool Fiberglass Production'' is one of the 174 source categories listed in the notice. As defined in the EPA report, ``Documentation for Developing the Initial Source Category List'' (docket item II-A-5), the Wool Fiberglass Production source category includes any facility engaged in producing wool fiberglass from sand, feldspar, sodium sulfate, anhydrous borax, boric acid, or any other materials. Facilities that manufacture mineral wool from rock, slag, and other similar materials are not included in the source category. A separate MACT standard for mineral wool production is currently under development. Before this project began, no formaldehyde test methods and no HAP data were available to assess the effectiveness of control devices in this industry for controlling HAP emissions. The EPA and the wool fiberglass industry worked in a partnership to address the data needs for the purpose of establishing a MACT standard. Through a cooperative effort, EPA and NAIMA developed methods for measuring formaldehyde emissions from wool fiberglass manufacturing processes. Using information supplied voluntarily by industry for each wool fiberglass manufacturing line, EPA identified processes and control systems as candidates for emissions testing that were considered representative of the MACT floor and MACT for new sources. EPA and the industry were able to obtain the necessary emissions data as a result of these cooperative efforts. Based on the information collected, EPA believes it is likely that all but three wool fiberglass plants are major sources subject to the proposed NESHAP. A major source must have the potential to emit 9.1 Mg/ yr (10 ton/yr) or more of a single HAP or 23 Mg/yr (25 ton/yr) or more of a combination of HAPs. Three facilities (each with one line producing bonded products) may be area sources. At these sites, two of the three glass-melting furnaces and all three RS forming processes are controlled at the MACT floor level. Because these facilities are not believed to present an adverse environmental or health risk, EPA has determined that it is not necessary to include these wool fiberglass manufacturing facilities on the list of area sources required by section 112(c)(3) of the Act. On December 3, 1993 (58 FR 63941), EPA published a schedule for the promulgation of standards for the sources selected for regulation under section 112(c) of the Act. According to this schedule, MACT standards for this source category must be promulgated no later than November 15, 1997. If standards are not promulgated by May 15, 1999 (18 months following the promulgation deadline), section 112(j) of the Act requires States or local agencies with approved permit programs to issue permits or revise existing permits containing either an equivalent emission limitation or an alternate emission limitation for HAP control. (See ``Guidelines for MACT Determinations Under Section 112(j),'' EPA 453/R-94-026, May 1994.) B. Selection of Emission Sources The wool fiberglass manufacturing source category, as defined in the EPA report, ``Documentation for Developing the Initial Source Category List,'' includes, but is not limited to: (1) The glass-melting furnace, (2) marble forming, (3) refining unit, (4) fiber formation process, (5) binder application process, (6) curing process, and (7) cooling process. For the reasons described below, EPA selected the forming, curing, and cooling processes on new and existing RS manufacturing lines and the forming and curing processes on existing FA manufacturing lines producing pipe insulation and on new FA manufacturing lines producing pipe insulation or heavy-density products for control under the proposed NESHAP. The proposed NESHAP also covers glass-melting furnaces located at wool fiberglass manufacturing facilities. Glass-melting furnaces are generally large, shallow, and well- insulated vessels that are heated from above by gas burners or from within by electrical current. About 66 percent of the glass-melting furnaces used in the wool fiberglass industry are all-electric, about 25 percent are gas-fired and about 9 percent are a combination of gas and electric. Glass pull rates for furnaces range from 18 to 272 Mg/d (20 to 300 ton/d). In the glass-melting furnaces, raw materials (e.g., sand, feldspar, sodium sulfate, anhydrous borax, boric acid) are introduced continuously or in batches on top of a bed of molten glass, where they mix and dissolve at temperatures ranging from 1,500 deg.C to 17,00 deg.C (2,700 deg.F to 3,100 deg.F), and are transformed by a series of chemical reactions to molten glass. Particulate emissions are caused by entrainment of dust from batch dumping and the combustion process and from volatilization of raw materials. Emissions of chromium result from entrainment of materials eroded from the refractory lining of the furnace and the furnace exhaust stack. Lead and arsenic are released from the batch materials and from the use of contaminated cullet (crushed recycled glass). Glass-melting furnaces may be either gas- fired, electric, or a combination of gas and electric. Emissions from glass-melting furnaces are typically controlled by baghouses or dry ESPs. One type of electric furnace, the cold top electric furnace, has low PM emissions without add-on controls as a result of its design. Operators of these units maintain a thick crust of raw materials on top of the molten glass, which impedes the release of heat and keeps the air temperature above the molten glass at or below 120 deg.C (250 deg.F). One of two methods may be used for the next stage of the process, fiber formation. In an RS forming process, a regulated flow of molten glass enters the center of a rotating spinner. Spinners are in a linear arrangement, with 2 to 12 spinners on a single line. Centrifugal action forces the molten glass out of the [[Page 15239]] spinners through hundreds of small orifices in the spinner wall to form glass threads. As the threads exit the spinner, a high-velocity air jet or a mixture of air and natural gas flame forces the threads downward, which attenuates the threads to form glass fibers. In the FA forming process, also known as the ``pot and marble'' process, glass marbles that were produced at separate on- or offsite facilities are fed into ceramic pots (typically 6 to 28 pots per line) that are heated to a high temperature. Glass strands flow by gravity down through holes in the bottom of the pot and are directed by pinch rollers. Following the pinch rollers, a high-velocity, high-temperature mixture of air and gas flame is used to attenuate the fibers. Particulate and organic emissions are released during the fiber-forming process due to volatilization of raw materials and entrainment of fiberglass particles in the process air stream. After the fibers are formed, they are sprayed with a binder. A typical binder consists of phenol-formaldehyde resin, water, urea, lignin, silane, and ammonia. The binder composition used in the RS and FA forming process is similar. Air, at a flow rate ranging from about 430 to 5,100 actual cubic meters per minute (15,000 to 180,000 acfm), forces the fibers downward onto a continuously moving conveyor to form a mat, which is conveyed to the curing oven. Emissions of formaldehyde, phenol, and methanol occur as a result of the vaporization of the volatile binder as it comes in contact with hot fibers and as a result of binder that is not deposited on the mat but passes through the conveyor and is exhausted to the atmosphere. HAP emissions from forming are controlled by process modifications, such as resin and binder chemistry and fiberization technology. The curing oven drives off moisture remaining on the fibers and sets the binder. The temperature of the curing oven varies for each product, ranging from about 180 deg.C to 320 deg.C (350 deg.F to 600 deg.F). Fans are used to draw hot air through the mat within each of the oven zones; the hot air may be recycled within each zone to conserve energy. The total air flow exiting the oven ranges from about 200 to 850 actual cubic meters per minute (7,000 to 30,000 acfm) for the RS process and from 85 to 480 actual cubic meters per minute (3,000 to 17,000 acfm) for the FA process. Emissions of formaldehyde, phenol, and methanol are the result of vaporization of volatile compounds in the binder. Emissions from about one-third of the curing ovens on RS manufacturing lines are controlled by thermal incinerators; the remainder are uncontrolled for organic HAP emissions. None of the curing ovens on FA manufacturing lines are controlled for organic HAPs. The quantity of binder solids sprayed onto the glass fibers is governed by the type of product being manufactured. Typically, about 70 percent of the binder applied to the fiberglass remains on the product. The remainder remains on the conveyor and is recycled back into the process via the wash water or is exhausted with the forming or curing oven air. Quality control checks are routinely performed to determine the product LOI, which ensures that the correct weight percent of binder is present in the product. After curing, the fiber mat is conveyed to a cooling section, where ambient air is forced through the mat to eliminate ``hot spots'' in the product and to facilitate finishing and packaging. Cooling air flow rates range from 140 to 990 actual cubic meters per minute (5,000 to 35,000 acfm). By the time the mat with its thermally set binder reaches cooling, emissions of formaldehyde, phenol, and methanol are relatively small compared to forming and curing. Cooling processes are not controlled for HAP emissions. Most FA manufacturing lines do not have cooling sections because the product is able to cool adequately between exiting the curing oven and reaching the finishing and handling sections. At the current level of control, existing glass-melting furnaces emit approximately 270 kg/yr (600 lb/yr) of HAP and 750 Mg/yr (830 ton/ yr) of PM. Under the proposed NESHAP, EPA expects that seven currently uncontrolled electric furnaces would install controls. Electric furnaces (excluding cold top electric furnaces) emit an estimated 9 kg/ yr (20 lb/yr) of HAP and about 635 Mg/yr (700 ton/yr) of PM. Control of these furnaces would ensure that all furnaces are controlled to the MACT floor emission level. Existing cold top electric furnaces (air temperature above the molten glass of 120 deg.C [250 deg.F] or less) are not equipped with add-on control devices. Particulate emissions from the 12 existing cold top electric furnaces are limited by the thick crust maintained on the molten glass surface. Emissions are estimated to be 27 kg/yr (60 lb/yr) of HAP and about 55 Mg/yr (60 ton/yr) of PM. These furnaces are expected to comply with the proposed emission limit without the need for add-on control devices. The EPA considered requiring controls for cold top electric furnaces and has determined that the cost effectiveness of additional controls beyond the floor is not reasonable. Manufacture of wool fiberglass releases an estimated 1,770 Mg/yr (1,950 ton/yr) of formaldehyde from RS and FA manufacturing lines. The Agency selected forming, curing, and cooling processes on all new and existing RS manufacturing lines and forming and curing processes on existing FA manufacturing lines producing pipe insulation and new FA manufacturing lines producing pipe insulation or heavy-density products for control under the proposed NESHAP. Because no controls are currently used, the MACT floor is no control and because the cost effectiveness of additional controls beyond the floor is not reasonable, the Agency is not setting emission limits for existing FA manufacturing lines producing light-density, automotive, or heavy- density products or new FA manufacturing lines producing light-density or automotive products. Because no plants have equipped forming or curing processes on these manufacturing lines with emission controls, the MACT floor is no control. The EPA considered beyond-the-floor controls for both RS and FA manufacturing lines and has determined that the cost effectiveness of additional controls does not justify going beyond the floor. C. Selection of Pollutants The EPA proposes to regulate emissions of formaldehyde, a HAP and surrogate for phenol and methanol emissions, and PM emissions, a surrogate for metal HAP emissions. Formaldehyde, phenol, methanol, and the metal HAPs are included on the list of HAPs under section 112(b) of the Act and are emitted from wool fiberglass manufacturing sources. Formaldehyde is the only organic HAP emitted from the wool fiberglass industry that has been identified to be a potential carcinogen. EPA proposes to regulate emissions of formaldehyde, phenol, and methanol using formaldehyde as a surrogate measure for the proposed emission limits for manufacturing lines. Use of formaldehyde as a surrogate allows a single emission limit rather than individual emission limits for formaldehyde, phenol, and methanol (which would require separate measurements) because when the formaldehyde emission limit is met, phenol and methanol emissions will also be reduced. [[Page 15240]] D. Selection of Proposed Standards for Existing and New Sources 1. Background After EPA has identified the specific source categories or subcategories of major sources to regulate under section 112, MACT standards must be set for each category or subcategory. Section 112 establishes a minimum baseline or ``floor'' for standards. For new sources, the standards for a source category or subcategory cannot be less stringent than the emission control that is achieved in practice by the best-controlled similar source. [See section 112(d)(3).] The standards for existing sources can be less stringent than standards for new sources, but they cannot be less stringent than the average emission limitation achieved by the best-performing 12 percent of existing sources for categories and subcategories with 30 or more sources, or the best-performing five sources for categories or subcategories with fewer than 30 sources. After the floor has been determined for a new or existing source in a source category or subcategory, the Administrator must set MACT standards that are no less stringent than the floor. Such standards must then be met by all sources within the category or subcategory. In establishing the standards, EPA may distinguish among classes, types, and sizes of sources within a category or subcategory. [See section 112(d)(1).] The next step in establishing MACT standards is to investigate regulatory alternatives. With MACT standards, only alternatives at least as stringent as the floor may be selected. Information about the industry is analyzed to develop model plants for projecting national impacts, including HAP emission reduction levels and cost, energy, and secondary impacts. Regulatory alternatives (which may be different levels of emissions control, equal to or more stringent than the floor levels) are then evaluated to select the regulatory alternative that best reflects the appropriate MACT level. The selected alternative may be more stringent than the MACT floor, but the control level selected must be technologically achievable. The regulatory alternatives and emission limits selected for new and existing sources may be different because of different MACT floors. The Agency may consider going beyond the floor to require more stringent controls. Here, EPA considers the achievable emission reductions of HAPs (and possibly other pollutants that are co- controlled), cost and economic impacts, energy impacts, and other nonair environmental impacts. The objective is to achieve the maximum degree of emissions reduction without unreasonable economic or other impacts. [See section 112(d)(2).] Subcategorization within a source category may be considered when there is enough evidence to demonstrate clearly that there are significant differences among the subcategories. The EPA examined the processes, the process operations, and other factors to determine if separate classes of units, operations, or other criteria have an effect on air emissions or their controllability. The EPA considered developing subcategories of glass-melting furnaces on the basis of the energy sources used to convert the raw materials to molten glass and their emission potential. Glass-melting furnaces are typically either gas-fired, electric, or a combination of gas and electric. After examining PM emissions data for gas, electric, and combination gas and electric furnaces, EPA concluded that there is a large amount of variability in PM emissions regardless of energy source and that most furnaces are already well controlled by either ESPs or baghouses. Therefore, EPA decided not to develop subcategories of glass-melting furnaces. Wool fiberglass manufacturing lines can be classified by the type of forming process (RS and FA) used. Approximately 90 percent of the wool fiberglass manufactured by the RS forming process is building insulation, whereas the wool fiberglass manufactured by the FA forming process is specialty products, such as automotive or filtration products. Because of the type of products, the RS and FA forming process differ significantly in the way fibers are formed, production rates, air flow and energy expended per ton of product, application of process modifications, and the amount of binder applied to the wool fiberglass. As a result of these differences in manufacturing methodologies, levels of pollutant emissions, and application of controls (such as process modifications), EPA subcategorized manufacturing lines into those using the RS forming process (RS manufacturing lines) and those using the FA forming process (FA manufacturing lines). RS manufacturing lines consist of forming, curing, and cooling. FA manufacturing lines consist of forming and curing processes; cooling is not a distinct separate process on FA manufacturing lines. FA manufacturing lines can be further subcategorized by the type of specialty product made. The FA subcategories include light-density, heavy-density, automotive, and pipe insulation products. Each of these subcategories is characterized by a specific range of LOIs and densities, which gives each subcategory a different emission potential. Also, the control measures that can be used to reduce HAP emissions, for example, process modifications, are different for the FA subcategories. For all these reasons, the proposed standards have different emission limits for RS manufacturing lines and FA manufacturing lines and, within the FA subcategory, different emission limits for two FA subcategories. 2. Selection of Floor Technologies In establishing these proposed emission standards, the add-on or process control technology representative of the MACT floor was determined for each subcategory. In general, these determinations were made on the basis of the performances of the technologies as reported by emission test results. The technologies determined to be the MACT floors are those determined to be the median of the technologies that are representative of the best performing 12 percent of the sources (for which there are emissions data) where there are more than 30 sources in the subcategory or the best performing five sources (for which there are emissions data) where there are fewer than 30 sources. Of the 56 existing glass-melting furnaces, 12 are controlled by ESPs and 25 by baghouses (more than one furnace may be controlled by a single control device). PM emissions data are available for 18 furnaces. Because the number of furnaces is greater than 30, the MACT floor is represented by the average of the best performing 12 percent of the existing sources. Based on PM emissions data for the best performing 12 percent, baghouses and ESPs are equally effective in controlling PM emissions from glass-melting furnaces. Therefore, the MACT floor for existing glass-melting furnaces is represented by well- designed and operated baghouses and ESPs. An ESP representative of the MACT floor will have a specific collection area of 32 square meters per 1,000 actual cubic meters per hour (590 ft \2\/1,000 acfm); a baghouse representative of the MACT floor is a pulse-jet baghouse with polyester bag material and an air-to-cloth ratio of 0.9 actual cubic meters per minute per square meter (3 acfm/ft \2\ ). Because the same well- designed and -operated baghouses and ESPs are considered by EPA to be the best control technology for PM emissions, including metal HAP emissions, MACT for new furnaces [[Page 15241]] would be the same as the MACT floor for existing sources, a baghouse or an ESP. HAP emissions control on RS forming processes is achieved by process modifications including resin and binder chemistry, fiberization technology, binder application, and forming conditions (docket item II-D-62). Resins are manufactured by an outside supplier or in-house using proprietary technologies to meet the specifications of the wool fiberglass manufacturer. Variables, such as the phenol-to- formaldehyde mole ratio, resin cook procedures, and catalysts, control both the free-formaldehyde and phenol levels as well as the types and relative percentage of phenol oligomers, all of which influence the levels of emissions and acceptability of a resin for a given process. Resin purchase specifications are typically written so that the free- formaldehyde content is ``not to exceed'' a certain level. In binder chemistry, the addition of various additives can reduce formaldehyde emissions. Urea, for example, added to the binder solution reacts with free formaldehyde, which can form stable, nonreversible urea formaldehyde compounds. In fiberization technology, temperature of the fiber veil is a critical process variable (a lower temperature may reduce HAP volatilization) affected by the fiberizer design and operation as well as by air and water treatment of the fiber veil. Binder application efficiency, the amount of binder that stays on the fiberglass, is increased by matching binder droplet size to the fiber diameter. Factors such as nozzle size geometry, configuration of the nozzle assembly, and location affect binder droplet size. Forming conditions, such as air volume and velocity affect binder application efficiency; too much or too little air flow can increase emissions. Each of these process modifications has been implemented on each of the 40 RS forming processes, although the degree to which each process modification has been implemented is different for each line. Add-on controls such as wet scrubbers or wet ESPs, primarily for PM control, were shown to be ineffective for gaseous HAP removal. Thus, the MACT floor for forming on existing RS manufacturing lines is represented by process modifications. Because the number of RS forming sources, 40, is greater than 30, the MACT floor is represented by the median of the best performing 12 percent of existing sources, or five sources (40x0.12=4.8). Based on HAP emissions data for the best performing 12 percent of existing sources, process modifications are the MACT floor for forming processes on RS manufacturing lines. Because of differences in application between companies and because of the proprietary nature of process modifications, a detailed description of forming process modifications cannot be presented. Of the 43 curing ovens on RS manufacturing lines, 14 are controlled using incinerators. Based on the median of the top 12 percent, the thermal incinerator is the MACT floor for curing processes on existing RS manufacturing lines. Thermal incinerators have been shown to be highly effective in the control of emissions of organic HAPs and can achieve destruction efficiencies in excess of 98 percent with an adequately high temperature, good mixing, sufficient oxygen, and adequate residence time. Low organic concentration gas streams, such as those emitted from wool fiberglass curing processes, can be expected to have low heating values and require auxiliary fuel. Heat recovery through the use of a recuperative incinerator can reduce the energy requirements. Emission test measurements demonstrate that a thermal incinerator is at least 99 percent effective in the removal of formaldehyde and phenol from curing ovens. Based on the median of the best performing 12 percent of existing sources, a thermal incinerator representative of the MACT floor has a combustion temperature of 700 deg.C (1,300 deg.F) and a gas residence time of 1 second. While the MACT floor for cooling is no control, cooling is included in the definition of RS manufacturing line, and therefore covered as part of the proposed RS manufacturing line standard. This inclusion prevents the shifting of emissions from forming and curing to the cooling section. The EPA's analysis of MACT floor control options for existing RS manufacturing lines (described above) showed that the median of the best performing 12 percent of existing forming processes control HAP emissions using process modifications and the median of the best performing 12 percent of existing curing ovens are controlled by incinerators. As a result, the MACT floor for RS manufacturing lines is forming process modifications coupled with an incinerator for curing emissions. These controls were determined to be the most efficient for the control of HAPs among the various controls used in the industry for existing RS manufacturing lines. Based on the best controlled source, MACT for new RS manufacturing lines is more stringent than the MACT floor for existing RS manufacturing lines. MACT for new RS forming processes incorporates a higher degree of process modifications than is present on most existing forming processes but which is available to all the industry and can be designed into new forming processes. Because the MACT floor for existing curing ovens, incinerators operating at 700 deg.C (1,300 deg.F) and a gas residence time of 1 second, represent the best-controlled source, MACT for new curing ovens is the same as the MACT floor for existing curing ovens. None of the cooling processes are controlled for gaseous HAPs; as a result, MACT for new cooling processes is no control. Thus, EPA has determined that the MACT floor for new RS manufacturing lines is represented by a high level of process modifications on RS forming processes, incinerators on curing ovens, and no control on cooling processes. As discussed earlier, none of the forming processes on FA manufacturing lines producing light-density or automotive products are equipped with HAP emission controls. Thus, the MACT floor is no control for forming processes on new and existing FA lines producing these products. The median of the best performing five lines (fewer than 30 sources) producing heavy-density products was determined to be no control; thus, the MACT floor for forming on existing FA manufacturing lines producing heavy-density products is no control. The best- controlled heavy-density forming process uses process modifications; therefore, process modifications are the basis for the MACT floor for the forming process on new FA manufacturing lines producing heavy- density products. Emissions from the forming process on all FA manufacturing lines producing pipe insulation are controlled by the same level of process modifications. Therefore, process modifications are the basis for the MACT floor for the forming process on all new and existing FA manufacturing lines producing pipe insulation. No control systems have been applied for the control of HAP emissions from curing ovens on FA manufacturing lines. Therefore, the MACT floor for curing ovens on new and existing FA manufacturing lines is no control. Although the MACT floor for curing is no control, curing is included in the definition of FA manufacturing line and, therefore, is covered as part of the proposed FA manufacturing line standard. This inclusion prevents the shifting of emissions from forming to the curing section. The EPA's analysis of MACT floor control options for existing FA [[Page 15242]] manufacturing lines producing pipe product showed the best performing five forming processes (fewer than 30 sources) controlled by the same level of process modifications and curing ovens uncontrolled for HAP emissions. As a result, the MACT floor for existing FA manufacturing lines producing pipe products is process modifications for forming and no control for curing. Because the same level of process modifications is used on forming processes on all FA manufacturing lines producing pipe products and because no HAP controls are used on curing ovens, EPA has determined that the MACT floor for new FA manufacturing lines producing pipe products is the same as the MACT floor for existing sources. As described above, the MACT floor for forming processes and curing ovens on existing FA manufacturing lines producing heavy-density products is no control; therefore, the MACT floor for existing FA manufacturing lines producing heavy-density products is no control. Based on the best-controlled source, MACT for new FA manufacturing lines producing heavy-density products is process modifications on forming. Because no curing ovens are controlled, the MACT floor for new curing ovens is no control, the same as the MACT floor for existing curing ovens. Thus, EPA has determined that the MACT floor for new FA manufacturing lines that produce heavy-density products is represented by process modifications on forming and no control on curing ovens. The EPA considered requiring controls beyond the MACT floor for glass-melting furnaces and RS and FA manufacturing lines. However, based on an assessment of the impacts of beyond-the-floor controls, EPA concluded that the cost effectiveness of an incremental reduction in emissions would make additional controls unreasonable (docket items II- A-12, II-B-17, II-B-22). 3. Emission Limits As part of this rulemaking, emissions data were collected from tests at 10 wool fiberglass plants and from other test data supplied by NAIMA to characterize uncontrolled and controlled emissions from the various processes and evaluate the effectiveness of existing control systems. Sites tested during this rulemaking were selected based on their use of the control technology identified as candidates for MACT floor. Using the test data, EPA established the MACT floor emission limits for existing and new sources. Emissions data were evaluated for 18 furnaces controlled by baghouses and ESPs (docket item II-I-20). Emissions ranged widely for both gas and electric furnaces and for both well-designed and well- operated baghouses and ESPs. Controlled PM emissions from all furnaces ranged from 0.01 to 0.54 kg/Mg (0.02 to 1.08 lb/ton) of glass pulled. Emissions of PM from baghouse-controlled furnaces ranged from 0.01 to 0.54 kg/Mg (0.02 to 1.08 lb/ton) of glass pulled and from 0.01 to 0.25 kg/Mg (0.02 to 0.5 lb/ton) of glass pulled for ESP-controlled furnaces. Controlled electric furnace PM emissions ranged from 0.01 to 0.35 kg/Mg (0.02 to 0.7 lb/ton) of glass pulled; controlled gas furnace emissions ranged from 0.01 to 0.54 kg/Mg (0.02 to 1.08 lb/ton). In proposing emission limits, EPA took into consideration the wide variation in controlled emissions for both gas and electric furnaces and for well- designed and operated baghouses and ESPs. The proposed PM emission limits represent a level that can be achieved by all existing furnaces that are controlled by well-designed and operated baghouses and ESPs. Because MACT for new and existing furnaces is the same, EPA proposed the same PM emission limit, 0.25 kg of PM/Mg (0.5 lb of PM/ton) of glass pulled, for new furnaces as for existing furnaces. The proposed PM emission limit for existing glass-melting furnaces, 0.25 kg/Mg (0.5 lb/ton) of glass pulled, is the same as the current NSPS level for gas- fired glass-melting furnaces in the wool fiberglass industry (see 40 CFR part 60, subpart CC). Both baghouses and ESPs are used to control emissions from gas-fired furnaces. In proposing the same emission limit for new and existing furnaces, EPA recognizes that both baghouses and ESPs used on existing furnaces are already highly efficient at controlling PM emissions and there is no basis for a more stringent emission limit based on this control technology. The limited emission test data for metal HAPs show their emissions to be low, often below the detection limits of the test method. In cooperative efforts by EPA and NAIMA, tests for metal HAPs were performed at six glass-melting furnaces (docket item II-B-15). For a medium capacity controlled furnace (27,000 Mg/yr [30,000 ton/yr]), emissions of arsenic would be 0.2 lb/yr, chromium emissions would range from 1.2 to 18 lb/yr, and lead emissions would be 0.6 to 2.1 lb/yr. Total metal HAP emissions from a large (50,000 Mg/yr [55,000 ton/yr]) controlled model gas-fired furnace are an estimated 60 lb/yr. For RS forming processes, the number of sources is 40. Because the number of sources is greater than 30, the MACT floor is represented by the median of the best performing 12 percent of existing sources, or five sources. Emissions of formaldehyde from forming processes representative of the best performing five were measured (docket items II-B-15, II-B-21, II-D-64). Emissions of formaldehyde from these five forming processes were 0.15, 0.33, 0.49, 0.49, and 0.6 kg/Mg (0.3, 0.65, 0.97, 0.97, and 1.2 lb/ton) of glass pulled. Using these results, the median emission level is 0.49 kg of formaldehyde per megagram (0.97 lb of formaldehyde per ton) of glass pulled. The emission level selected as representative of new forming processes, 0.33 kg of formaldehyde per megagram (0.65 lb of formaldehyde per ton) of glass pulled, reflects the performance of the best process modification available to the industry. The emission level of 0.15 kg/Mg (0.3 lb/ ton) is from a proprietary forming process not available to the rest of the industry. Therefore, it was not considered MACT for new sources. Emissions test results for RS forming processes are summarized in Table 4. Table 4.--Summary of Emission Test Results on RS Manufacturing Lines [Docket Items II-B-15, II-B-21, II-D-64] ---------------------------------------------------------------------------------------------------------------- Average Formaldehyde Emissions Process and Plant Control ----------------------- kg/mg lb/ton ---------------------------------------------------------------------------------------------------------------- Forming Process modificationsa Plant P........................................... .................................... 0.15 0.3 Plant S........................................... .................................... 0.33 0.65 Plant T........................................... .................................... 0.6 1.2 Plant U........................................... .................................... 0.49 0.97 [[Page 15243]] Plant V........................................... .................................... 0.49 0.97 Curing Plant M........................................... Incinerator (1300 deg.F, 0.5-s residence time) Inlet............................. 0.497 0.994 Outlet............................ 0.00039 0.00078 Plant N........................................... Incinerator (1500 deg.F, 2.5-s residence time) Outlet............................ 0.00146 0.00292 Cooling Plant O........................................... Uncontrolled........................ 0.004 0.007 ---------------------------------------------------------------------------------------------------------------- a Process modifications include resin chemistry, binder chemistry, fiberization technology, binder application, forming conditions. RS curing processes, controlled by incinerators, were tested at two plants using the technology that EPA determined represented the MACT floor for RS curing, resulting in one measurement of 0.0004 kg of formaldehyde per megagram (0.001 lb of formaldehyde per ton) of glass pulled and another measurement of 0.0015 kg of formaldehyde per megagram (0.003 lb of formaldehyde per ton) of glass pulled (docket item II-B-15). Because results from just two tests were available, the higher result (0.0015 kg of formaldehyde per megagram [0.003 lb of formaldehyde per ton] of glass pulled) was chosen to represent MACT floor emissions from existing and new curing ovens. The only test result for emissions from cooling operations was 0.005 kg of formaldehyde per megagram (0.01 lb of formaldehyde per ton) of glass pulled (docket item II-B-15); this emission level was selected to represent the emissions from new and existing cooling processes. Emissions data for RS curing and cooling processes are summarized in Table 4. The proposed formaldehyde emission limit for existing RS manufacturing lines, 0.6 kg of formaldehyde per megagram (1.2 lb of formaldehyde per ton) of glass pulled, is based on the combined manufacturing line emission levels from forming, curing, and cooling with a 20 percent allowance to account for the use of short-term test data as compared to long-term continuous monitoring data. In metric units, the emission limit for existing RS manufacturing lines was calculated as follows: (0.49 + 0.0015 + 0.005) x 1.20 = 0.6 kg of formaldehyde per megagram of glass pulled. In English units, the emission limit for existing RS manufacturing lines was calculated as follows: (0.97 + 0.003 + 0.01) x 1.20 = 1.2 lb of formaldehyde per ton of glass pulled. The proposed emission limit for new RS manufacturing lines, 0.4 kg of formaldehyde per megagram (0.8 lb of formaldehyde per ton) of glass pulled, was derived using 0.33 kg/Mg (0.65 lb/ton) for the forming emission level and the same emission levels for curing and cooling as mentioned above. In metric units, the emission limit for new RS manufacturing lines was calculated as follows: (0.33 + 0.0015 + 0.005) x 1.20 = 0.4 kg of formaldehyde per megagram of glass pulled. In English units, the emission limit for new RS manufacturing lines was calculated as follows: (0.65 + 0.003 + 0.01) x 1.20 = 0.8 lb of formaldehyde per ton of glass pulled. For existing and new FA manufacturing lines that produce pipe insulation, the MACT floor for forming is the same process modification, which has been applied to an equal degree to all forming processes. Because there are no formaldehyde emission controls on curing on FA manufacturing lines producing pipe insulation, the MACT floor for curing is no control. Emissions of formaldehyde have been measured from forming and curing on six FA manufacturing lines producing pipe insulation where the same MACT floors for forming and curing were used (see Table 5). Results from short-term formaldehyde emission tests on these FA manufacturing lines were 1.7, 2.4, 2.4, 2.4, 3.2 and 3.4 kg/Mg (3.4, 4.7, 4.8, 4.9, 6.5, and 6.8 lb/ton) of glass pulled (docket item II-D-54). Even though the same control technologies and methods on manufacturing lines (forming and curing) producing the same product were used, the emissions varied widely from 3.4 to 6.8 lb/ ton. Because the test data for the same control technologies and methods that represent the MACT floors show a range of emissions and because emissions tests used short term tests (3 hrs) while the MACT standard will need to be met at all times, EPA has set the proposed formaldehyde emission limit for new and existing FA manufacturing lines producing pipe insulation at 3.4 kg of formaldehyde per megagram (6.8 lb of formaldehyde per ton) of glass pulled. The EPA believes that this emission rate is the level that can be consistently achieved by the control technologies and methods that are the MACT floor. Table 5.--Summary of Emissions Data for FA Manufacturing Lines [Docket item II-D-54] ------------------------------------------------------------------------ Formaldehyde emissions Process and product Control ---------------------- kg/mg lb/ton ------------------------------------------------------------------------ Heavy density.............. Forming--process 2.3 4.6 modifications. 3.9 7.8 Curing--no control.. [[Page 15244]] Pipe....................... Forming--process 1.7 3.4 modifications. 2.35 4.7 Curing--no control.. 2.4 4.8 2.45 4.9 3.25 6.5 3.4 6.8 ------------------------------------------------------------------------ In the case of new FA manufacturing lines that produce heavy- density product, the MACT floor is represented by process modifications on forming processes, which have been applied to the same degree on two forming processes, and no control on curing. The emission limit selected for new FA manufacturing lines producing heavy-density product is based on the results of emissions testing on forming and curing processes on two FA manufacturing lines producing heavy-density products where the same process modifications have been applied to forming and both curing ovens are uncontrolled (see Table 5). Emissions of formaldehyde from these two FA manufacturing lines were 2.3 and 3.9 kg of formaldehyde per megagram (4.6 and 7.8 lb of formaldehyde per ton) of glass pulled (docket item II-D-54). Because of the small number of tests, the use of short-term test data (rather than long-term continuous monitoring data), and to allow for the variability in emission results from forming processes using the same floor level process modifications, the 3.9 kg/Mg (7.8 lb/ton) level was chosen to represent MACT floor emissions from new FA manufacturing lines manufacturing heavy-density products. E. Selection of Monitoring Requirements Several monitoring options were identified and evaluated for sources in wool fiberglass manufacturing facilities. Under the most stringent option, a continuous opacity monitor (COM) would be required for monitoring PM emissions from glass-melting furnaces, and a continuous emission monitor (CEM) would be required for measurements of formaldehyde, phenol, and methanol. No EPA-approved continuous monitoring method is available for measuring PM, which is used as a surrogate for metal HAP emissions. Where continuous monitors do not exist or are too expensive, monitoring would rely on parametric monitoring of one or more parameters associated with the production process or control device, coupled with corrective action for operating problems. Potential parameters could include incinerator operating temperature, ESP electrical readings, and binder formulation parameters. A bag leak detection system could be used to monitor PM emissions from baghouses and ensure proper operation and maintenance of the control devices. Visible emissions observations by Method 9 could be required on a daily or weekly basis to ensure proper operation of control devices on glass- melting furnaces. For this industry, however, opacity is not considered a good indicator of compliance because of the low grain loadings. Therefore, this option was not considered further. A one-time performance test is necessary to demonstrate compliance with the applicable emission limit for glass-melting furnaces and manufacturing lines. Using the surrogate approach, the owner or operator would measure PM emissions from the furnace control system using EPA Method 5 in appendix A to 40 CFR part 60 and Sec. 63.1389 (Test methods and procedures) and formaldehyde emissions using EPA Method 316 or Method 318. Methods 316 and 318 are also being proposed today. The sampling and analytical cost for a three-run performance test is estimated at $8,000 for Method 5 and $9,000 for Method 316. The owner or operator could also use EPA Method 318, for measuring formaldehyde emissions for compliance purposes as well measuring other pollutant emissions. The method is also validated for use as a CEM. The sampling and analytical cost for three Fourier Transform Infrared (FTIR) gas-phase extractive runs, including other tests needed in conjunction with Method 318, is about $15,000. During the performance tests for each glass-melting furnace and each RS and FA manufacturing line subject to the standard, the owner or operator would monitor and record the glass pull rate and determine the arithmetic mean for each test run. A determination of compliance during the performance tests would be based on the average of the three individual test runs. Each owner or operator subject to the proposed NESHAP would submit a written operations, maintenance, and monitoring plan as part of their application for a part 70 permit. The plan would include procedures for the proper operation and maintenance of processes and add-on control devices used to comply with the proposed emission limits as well as the corrective actions to be taken when a process or control device parameter deviates from allowable levels established during performance testing. The plan would identify the process parameters and control device parameters that would be monitored to determine compliance, a monitoring schedule, and procedures for keeping records to document compliance. Additional information may be required depending on the add-on control device or process that is used to comply with the emission standard. The owner or operator of each furnace controlled by an ESP would submit as part of their operations, maintenance, and monitoring plan the ESP parameters (e.g., secondary voltage of each electrical field) to be monitored, a monitoring schedule, recordkeeping procedures to document compliance, and how the ESP is to be maintained and operated. The proposed monitoring provisions specify that corrective actions be taken according to the procedures in the operations, maintenance, and monitoring plan in the event of a deviation in any 3-hour average ESP parameter outside the range established during performance testing. Failure to initiate corrective actions within 1 hour of the deviation would be considered noncompliance. If the ESP [[Page 15245]] parameter values are outside the range established during the performance test for more than 5 percent of total operating time in a 6-month reporting period, the owner or operator would implement a QIP consistent with subpart D of the draft approach to compliance assurance monitoring.7 If the ESP parameter values are outside the range for more than 10 percent of total operating time in a 6-month reporting period, the owner or operator would be in violation of the standard. --------------------------------------------------------------------------- \7\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- Following the performance test, the owner or operator of each glass-melting furnace controlled by a baghouse would monitor emissions exiting the PM control system using a bag leak detection system since opacity is not a good indicator of performance at the low, controlled PM levels characteristic of these sources. The bag leak detection system must be equipped with an alarm system that will sound when an increase in PM emissions is detected. On a positive pressure baghouse where more than a single bag leak detection system probe may be necessary, the instrumentation and alarm for the bag leak detection system may be shared among detectors. Provisions are included in the rule regarding installation, calibration, and operation of the system. The monitoring provisions specify that when the bag leak detection system alarm is activated, the baghouse be inspected for the cause of the alarm and that corrective action be initiated according to the procedures in the operations, maintenance, and monitoring plan. Failure to initiate corrective actions within 1 hour of the alarm would be considered noncompliance. If the alarm is activated for more than 5 percent of the total operating time during the 6-month reporting period, the owner or operator must implement a QIP consistent with subpart D of the draft approach to compliance assurance monitoring.8 --------------------------------------------------------------------------- \8\ Proposed rule published in the August 13, 1996 Federal Register (61 FR 41991). --------------------------------------------------------------------------- The owner or operator of a gl [Text truncated at 120,000 characters. The full text is on the page linked above.] --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A97-7214. Public record. Not legal advice.