Document oGLDQr963wmggjeXOYkJVmr8

ER^PjET April 28, 2017 VIA OVERNIGHT MAIL AMP EMAIL Honorable Scott Pruitt, Administrator U.S, Environmental Protection Agency USEPA Headquarters William Jefferson Clinton Building 1200 Pennsylvania Avenue, N.W. Room 3000 Washington, DC 20004 Mr. Steffan Johnson Leader, Measurement Technology Group Office of Air Quality Planning and Standards U.S. Environmental Protection Agency (Mail Code El 43-02) Research Triangle Park, NC 27711 johttsoii.SteffaitfSiepa.goy Re: Request for Approval of Alternative to Test Method; National Emission Standards for Hazardous Air Pollutants: Ferroalloys Production, 82 Fed. Rea. 5401 (January 18, 2017) Dear Administrator Pruitt: Pursuant to 40 C.F.R. 63.7(e)(2)(ii), Eramet Marietta Inc. ('`EMI") hereby respectfully requests the Administratoi's approval of an alternative to the performance test method specified at 40 C.F.R. 63.1625(d)(1) for conducting shop building opacity observations at EMI's existing ferroalloy production facility in Marietta, Ohio. As explained in greater detail below, EMI requests approval to use EPA Test Method 9 (as set forth in 40 C.F.R. Part 60 Appendix A-4) in lieu of the ASTM D7520-16. IilEiiicijti EMI is subject to the National Emissions Standards for Hazardous Air Pollutants: Ferroalloys Production ("Ferroalloys NESHAP''). 40 C.F.R. Subpart XXX, promulgated at 80 Fed. Reg. 37366 (June 30, 2015), as amended at 82 Fed. Reg. 5401 (January 18, 2017). Under 40 C.F.R 63.1625(d)(1) of the Ferroalloys NES11AP. compliance with conducting shop building opacity standards is to be demonstrated using ASTM D7520-16. also referred to by EPA as the Digital Camera Opacity Technique or "DCOTT The applicable opacity standard is set forth at 40 C.F.R. 63.1623(b)(3): ' Unless you meet the criteria of 63.1623(b)(3)(i!i), you must not cause the emissions exiting from a shop building to exceed an average of 8 percent opacity over a furnace or MOR process cycle. ERAMET littftSST.Vfi-' KOVTE 7 SOOTH ttAlitOTA, OH 45M TE!.:VW.J"4 *000 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00001 (i) This 8 percent opacity requirement is determined by averaging the individual opacity readings observed during the furnace or MOR process cycle. (ii) An individual opacity reading shall be determined as the average of 24 consecutive images recorded at 15-second intervals with the opacity values from each individual digital image rounded to the nearest 5 percent. (iii) If the average opacity from the shop building is greater than 8 percent opacity during an observed furnace process cycle, the opacity of two more furnace process cycles must be observed within 7 days and the average of the individual opacity readings during the three observation periods must be less than 8 percent opacity. (iv) At no time during operation may the average of any two consecutive individual opacity readings be greater than 20 percent opacity. Under 40 C.F.R. 63.7(e)(2)(ii), the Administrator may approve the use of "an intermediate or major change or alternative to a test method (see definitions in 63.90(a)), the results of which the Administrator has determined to be adequate for indicating whether a specific affected source is in compliance...."' The request generally follows the framework prescribed in 40 C.F.R. 63.8(f)(4) for requests to use an "alternative test method." Under 63.8(t)(2). the Administrator may approve "Alternatives to the American Society for Testing and Materials (ASTM) test methods or sampling procedures specified by any relevant standard.. However, EMI notes that this framework is not strictly applicable to a request to use EPA's long-established Method 9, because KPA has already validated the requested method for use in conducting the sort opacity observations involved in this request. See 40 C.F.R. 63.2 {``"Alternative lest method means any method of sampling and analyzing for an air pollutant that is not a test method in this chapter.. A). As a consequence, certain aspects of the alternative lest method application - most notably, compliance with the alternative test method validation procedures of 40 C.F.R. 63.7(f) -- do not make sense in the context of this request to use a previously approved and widely used EPA test method. As EPA has stated in the context of a proposed alternative test method, the "principal criterion" is whether a proposed method is at least as stringent as the specified method: The EPA Administrator or his authorized representative must be satisfied that the test method alternative will produce results adequate to determine compliance. In other words, the EPA Administrator or authorized representative, such as a State having delegated authority, generally must be assured that a lest method change provides a determination of compliance status at the same or greater stringency as the test method specified in the applicable regulation.1 1 Since this request involves the use of an entirely different opacity test method ~ EPA Method 9 -- than that specified in the Ferroalloys NESHAP, it might also be treated as a "major change to test method," as defined in 40 C.F.R. 63.90(a). 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00002 72 Fed. Reg. 4257, 4261 (January 30, 2007).2 In this case, EPA lias already reaffirmed that Method 9, which thousands of facilities across the United States currently use to demonstrate compliance with opacity standards, and which EPA has previously authorized EMI to use to measure the precise emissions involved in this request, is as accurate as ASTM D7520-16 in measuring opacity. "[Wje believe, based on validation studies that EPA Method 9 and [ASTM D7520-16] provide comparable opacity results, , . A 82 Fed. Reg. 5401, 5404-05 (January 18, 2017);3 see also, e.g., "Summary of Public Comments and Responses on the Proposed Reconsideration (81 FR 45089, July 12, 2016) of the Final Rule (80 FR 37366, June 30, 2015)''' ("Response to Comments"), EPA-HQ-OAR-2010-0895-034!, at 6 (disagreeing with commenteds suggestion that "Method 9 is less accurate than DCOT" and stating that "both methods provide comparable measurements of opacity"). Indeed, EPA has treated Method 9 as the benchmark against which the accuracy of ASTM D7520-16 should be measured, describing ASTM 1)7520-16 as "an acceptable alternative to EPA Method 9. . . Id, at 5407. The Ferroalloys NESHAP actually provides for the use of EPA Method 9 to certify the accuracy of ASTM D7520-16. See 40 C.F.R. 63.1656(b)(7). Information in Support of This Request 63.8(f)(4)(i): Submit request at least 60 days before testing to demonstrate compliance with a standard. EMI requests approval to use Method 9 to demonstrate compliance with the opacity standard set forth at 40 C.F.R. 63.1623(b)(3), beginning June 30, 2017. This request is being submitted more than 60 days before the applicable testing will begin. 63.8(f)(4)(ii): Description ofproposed alternative monitoring system which addresses the four elements contained in the definition of monitoring in 63.2 and information justifying the requestfor an alternative monitoring method. A. Name, location, and affected units Eramet Marietta Inc. 16705 Stale Route 7 Marietta. OH45750-0299 2 See also "Requests for Approval of Aiteraatives/Modifications to Test Methods and Testing Procedures." EMC GD-022R4, Air Quality Assessment Division, OAQPS, EPA (April 21,2014) ("EMC Guidance"; at 1 (basic principles for review of request for major change to test method are (a) whether alternative method "will produce results adequate for the Administrator's determination of compliance" and (b) whether the change will "affect the stringency of the applicable regulation"). ' EPA previously mischaracterized EMI's August 25, 2015 petition for reconsideration of the requirement to use ASTM D7520-13 as a "request to allow' EPA Method 9 as an alternative method for determining compliance " 82 Fed. Reg. at 5405, EMI's petition was expressly submitted pursuant to Section 307(d) of the Clean Air Act, 42 li.S.C. 7607(d), and not as request for an alternative test method, whose provisions are not directly applicable to such a request. 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00003 B. Elements of Monitoring Under 63.2 1* Indicators) of performance EPA Method 9 provides direct measurement of opacity by trained and certified readers. 2. Measurement techniques EPA Method 9 uses direct visual observation of visible emissions opacity by trained and certified readers. 3. Monitoring frequency Consistent with the requirements of 40 C.F.R. 63.1623(b)(3), EPA Method 9 collects observations at 15-second intervals over the course of the prescribed monitoring cycle. 4. Averaging time Consistent with the requirements of 40 C.F.R. 63.1623(b)(3), EPA Method 9 bases individual opacity readings based on the average of 24 consecutive observations collected at 15-second intervals. C. Information Justifying the Request 1. Method 9 produces results adequate for the determination of compliance. Method 9 is a Category A method approved by EPA Air Emission Measurement Center, promulgated in the Federal Register and codified in the Code of Federal Regulations. Method 9 is a proven method and has been used since the mid-1970's at thousands of industrial facilities to demonstrate compliance with opacity standards established under state and federal air pollution law. EPA itself has asserted that Method 9 offers the same measurement capability as ASTM D7520 in the context of this rule: "We also agree that ASTM D7520 provides requirements adequate to read roof vents and those requirements are parallel to those of EPA Method 9. Therefore, there are no real differences in capability there Response to Comments at 5 (emphasis added). EPA recognized that Method 9 would be adequate for the determination of compliance when it asserted that Ohio EPA could use Method 9, and need not purchase DCOT to independently assess EMI's compliance with the ferroalloy NESHAP's opacity standards "[State Agencies] will also he able to conduct EPA Method 9 readings which can serve as Credible Evidence if there are issues with the opacity being reported," Id. at 11 (emphasis added). Similarly, Method 9 is already accepted by EPA Region 5 as an alternative to DCOT to demonstrate compliance in an enforcement context, as evidenced by the proposed consent decree lodged with the 13,5, District Court for the Eastern District of Wisconsin in the enforcement 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00004 action captioned, United States ofAmerica v. Maynard Steel Casting Company, Civil Action No. 2;17-cv-00292. Section 2.2 of Appendix C to the proposed consent decree expressly provides the respondent with the option of using Method 9 in lieu of DCOT. 2. Use of Method 9 will not affect the stringency of the applicable regulation.___________________________ 2.1. Ferroalloys NESHAP opacity standard was developed using Method 9 While EMI continues to believe that DCOT is not a reliable method to assess compliance with the roof vent opacity standards in the Ferroalloys NESHAP. EPA's own statements establish that Method 9 is at least as exact and rigorous as DCOT. Indeed, it would be impossible for Method 9 to decrease the stringency of the opacity standard in the Ferroalloys NESHAP, since that standard was developed, and presented in EPA's October 6, 2014 proposed rale, based on the express expectation that it would be enforced via Method 9. In that proposal, the opacity standard was designed to match Method 9's data collection protocol, with observations recorded at 15-second intervals (rounded to the nearest five percent) and averaged over six-minute blocks, and DCOT was presented only as an alternative to Method 9. Since EPA did not alter its Method 9-based opacity standard when replacing Method 9 with DCOT in the final rule, there is no rational basis for concluding that use of Method 9 would reduce the stringency of that standard. 2.2. Use of Method 9 does not affect opacity monitoring exactness DCOT has been developed to match the results of visual observation through Method 9. The digital camera used at a facility must be jointly certified with the analysis software by comparing DCOT results to visual observations made by certified Method 9 readers. See ASl'M D7520-] 6 paragraph 9.2. u[T]here is no basis that the Agency is aware offor the statement that EPA Method 9 readers are inaccurate after 6 minutes and the commenter did not provide references to support that claim. We believe that the EPA Method 9 collaborative studies demonstrate that readers can observe for significantly longer periods of time on several types of sources." Response to Comments at 5. 2.3. Use of Method 9 does not affect opacity monitoring rigor Method 9 procedures are applied by certified opacity readers. Certified readers are trained by Smoke schools. Smoke schools are quality controlled by EPA. Certified readers are trained every 6 months to keep their certification. 2.4. Use of Method 9 does not affect opacity monitoring objectivity DCOT imposition is based on a misconception, the belief that DCOT is more objective because if supposedly removes human subjectivity in determining opacity: "With respect to the comments requesting supporting data to support EPA A position that DCOT provides more objective and belter substantiated readings. We based this assessment on lhe fact that DCOT provides a record of every opacity observation and the opacity is measured using software algorithms to determine the contrast between the opacity and the background. We believe that 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00005 using software algorithms provides a consistent and objective determination oj opacity, and removes any bias by the observer...... Our position is that both methods provide comparable measurements of opacity, however DCOT provides a superior, objective record for its opacify measurements." Response to Comments at 6. DCOT does not, remove human intervention: ASTM D7520-16 identifies multiple points where human judgment (anti therefore possibilities for bias) is involved: "analysis softwaresoftware that when combined with a defined operating environment: (a) inputs images captured by the Digital Still Camera image capture devices: (b) produces opacity measurements from the combination of human interadkm, open or proprietary calculations and algorithms and image content viewing: (c) and then output said opacity measurement along with Analysis Software ',v configuration, image source documentation and other environmental parameter, " (Section 3.2.1) (emphasis added). "DCOT operator-refers to the human operating the DCOT system who records the digital still images with the Digital Still Camera and (hen determines plume opacity with the Analysis Software." (Section 3.2.4) (emphasis added). "A Digital Still Camera is used to capture a set of digital images of a plume against a contrasting background. Each image is analyzed with software that determines plume opacity by comparing a user defined portion ofthe plume image where opacity is being measured in comparison to the back ground providing the contrasting valuesC (Section 4.11 (emphasis added). DCOT seemed to be preferred by EPA for being less dependent on human decision and, therefore, less subjective. DCOT does not remove subjectivity: the DCOT operator, in its role as software analyst must decide which portion of the plume to choose on the pictures and which portion of the picture represents the background; therefore, the analyst has to make two subjective decisions where the Method 9 certified reader makes one. Moreover the DCOT model offered by Virtual Technology LLC does not eliminate intervention by an employee of the facility, which seems to be a motivation of EPA in imposing DCOT. The integrity of EMI's Method 9 Certified Readers has not been questioned in all the years of EMI's use of Method 9. In any event, the Virtual Technology model requires that the customer conducts QA/QC of all pictures and flags out pictures with errors, which maintains a high level of human intervention. 3. EMI is requesting the use of Method 9 for compelling reasons. 3.1. DCOT has significant deficiencies. Dcot is unproven for its intended use: DCOT is unproven for the use specified in the Ferroalloys NE5HAP. As described in detail in prior submissions, see EPA-HQ-OAR-20100895-320, 324, 326, 329. and 331-334, DCOT's suitability for the opacity measurements required in the Ferroalloys NESHAP was not validated by EPA following its own protocols. As a consequence, DCOT is being imposed on the manganese ferroalloys industry as a compliance test method although the technology and associated procedures did not undergo appropriate assessment during the rule development process. Dcot software is not fully developed for its intended use: DOCS II, the DCOT software sold by Virtual Technology LLC, the sole vendor in the market, is not fully developed and it is not ready to meet the requirements of the Ferroalloys NESHAP. ASTM D7520-16 stipulates in paragraph 7.1., "The DCOT system is formulated into three distinct and severable 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00006 components: (l) Digital Still Camera, (2) Analysis Software and Us associated computing platform, and {3) the output function. " Section 7.1.3, describes the output function. In an email dated March 14, 2017, Virtual Technology's Shawn Dolan confirmed that no Ferroalloys NESHAP output function exists m a certified version of DCOS IL In a correspondence dated April 18, 2017, in response to EMI's March 7, 2017 Request for Quotation ("RFQ"), Mr Dolan confirms that the output function includes the averaging calculations. He also writes: '"the report desiredfor the NESHAP referenced will have to he generated, tested and certified. This output creation and certification is a long lead time item ... as the entire system will require recertification to produce the desired output. A Note that in an earlier correspondence dated March 15 2017 in response to EMI's question about DCOS II readiness "can one extract from the analysis software a table with each single opacity value calculate averages"''. Mr Doian had answered:M7fR/M7, can extract into excel hut averages calculated outside of DCOS 11 are not DCOT Output and violate certification. e.g. "Output ' as defined in ASTM d?520~ 16,7.1.3' Dcot software development cannot be completed without detailed specifications from the REGULATOR: DCO l's sole supplier is not able, based on the NESHAP rules, to program this part of the software. As recently as April 18, 2017. Mr. Dolan wrote: "the current NESHAP does not define the averaging report well enough to create a final output function to support the report and certify the resultant configurationP He also wrote: "The specific report and its associated certification however i.s dependent on proper definition which is controlled by [EMI] and the regidalors associatedT 3-2. Meaningful improvements by using Method 9. Method 9 allows immediate corrective actions: Method 9 certified readers in the facility who arc also knowledgeable about the process, can have an immediate interaction with workers and supervisors operating the equipment, if an opacity observation indicates a trend toward exceeding the limits, quick action can be taken to resolve technical issues. This interaction is not achievable with a remote DCOT analyst that serves several customers, has no knowledge of the ferroalloys production process, and is not in real-time communication with operators. Dedicating a DCOT analyst at Virtual Technology to take charge of this interaction in real time would be costly (if achievable at all) and far less effective than using the Method 9 certified reader. Assuming DCOT was a proven technology for its intended use -- which EMI still disputes -- DCOT would be of no use to proactively respond to excessive fugitive emissions with a turnaround time of 45 minutes (time given by the sole vendor to EPA). EPA asserts in the Response to Comments: "The time estimate for performing and analyzing the images using DCOT was estimated by the vendor to take approximately 45 minutes depending on the contrasting background. We believe this is comparable to the amount of time required to calculate the 6-minutes average using EPA Method 9. The facility can always have a staffEPA Method 9 reader conduct informal readings and address problems in a timely manner if there are issues with the timeliness ofDCOT opacity results". Id. at 14. EPA's statement inviting the use of a Method 9 reader is both a recognition that Method 9 is adequate for opacity reading and that Method 9 has a faster turnaround time, EMPs own experience in Method 9 reading throughout 2016 shows that Method 9 readers have a far quicker turnaround time than 45 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00007 minutes. DCOT as it works today is merely a tool for sanctions with no benefit for the environment. Method 9 allows quick adjustments to unpredictable change in monitoring conditions: A method 9 certified reader can compensate for local weather issues such as change in wind velocity: such circumstances cannot be ruled out when observations are to be made over 60 to 120 minutes process cycles. A DCOT analyst cannot since he would base his weather assessment on data collected from the nearest National Weather Service Database in Parkersburg, Mid-Ohio Valley Regional Airport. 3-3. Method 9 provides a cost effective opacity monitoring method Method 9 is cost effective while DCOT is an unnecessary regulatory burden, without any real benefit to the environment, that if maintained would constitute an economic hindrance to the ferroalloys industry for the following reasons: Pool 's HIGH unpredictable INFRASTRUCTURE cost: DCOT's sole supplier asserted in an email to EMI dated January 26, 2016 that EMI could not use a digital camera on a tripod to conduct opacity testing under the Ferroalloys NESHAP. Therefore, multiple fixed cameras mounted on poles or buildings would have to be installed to cover EMI's three shop buildings covered by the rule. In light of the unreasonable fee Virtual Technology LLC intended to charge just to respond to EMI's RFQ. EMI undertook a preliminary survey with a specialist well- versed with Method 9 and digital camera technology. The survey helped in determining camera locations that would comply with requirements for slant angle, distance, and sun angle over each process cycle throughout each year's 365 days for each of its 3 process buildings. Assuming observations are limited to morning sun angles, the preliminary cost estimate is a minimum of $261,000, assuming a wired solution so as to secure a reliable electricity supply and fiber opticcable for reliable camera control and data transfer. The cost would be doubled if regulators require the capability to conduct observations during both the morning and afternoon. This cost was not reflected in the Ferroalloys NESHAP. and is not required with the use of Method 9 . Dcot's unpredictable implementation costs: EMI estimates that the cost to implement DCOT will be in the range of $410,000 per year if observations are limited to mornings, and $671,000 per year if regulators expect observations to be made mornings or afternoon alike. This estimate does not account for the time that EMFs employees would have to spend on the implementation work to coordinate the project. Likewise this estimate does not account for the cost of obtaining system certification per ASTM D7520-16, which is a complete unknown. This estimate is based on the price list given by Virtual Technology in July 2015. It is worth outlining that only after EPA published its final action on EMFs petition for reconsideration, did the sole DCOT supplier inform EMI that these prices no longer applied and that the vendor was no longer publishing a price list. Finally, these estimates are based on EMFs own preliminary evaluation, of the Output software development cost that will be priced by the supplier at $250/hour once the specifications have been dearly defined for the programming to stall. This cost is far in excess of what is required with the use of Method 9. Dcot's unpredictable operating COSTS: EPA's support for the use of DCOT Is based, at least in part, on a significant underestimate of its cost to implement. "With regard to the cost 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00008 ofDCOT we estimated in the final rule that the cast ofiImplementing the DCOT system would he approximately $200,000 per year for the source category with weekly readings. However these costs' decrease to about $90,000 per year for the source category if the facilities da monthly readings per furnace building. We believe that this is a reasonable cost to ensure the process fugitive emissions are effectively captured.'' Response to Comments at 17. EMI anticipates that the annual cost will far exceed EPA's estimate. EMI calculates that the operating costs for DCOT, if based on the 2015 Virtual Technology price list, will be in the range of a minimum of $186,000 per year assuming (i) weekly observations and (ii) no need to repeat the weekly tests due to adverse and unpredictable weather conditions that would interrupt an observation already started. If the observations were reduced to once per month the estimated cost is in the range of a minimum of $50,000 per year with the same assumptions. It is important to outline, however, that the DCOT sole vendor no longer publishes a price list and has indicated the 2015 price list is no longer applicable. This estimated annua! DC'Of cost does not include necessary EMI personnel costs. The EMI personnel costs to perform opacity readings will remain the same as for Method 9, since the time that EMI's employees will have to allocate to the collection of opacity images for DCOT is at least as high as when conducting Method 9 observations. 'The Virtual Technology DCOT model requires that the QA/QC responsibility remains with the customer, who must review all pictures and opacity results. See also ASTM D7520-16 in 3.2.6: "All of the digital images obtained by a DCOT system shall be reviewed by a qualified human DCOT operator to assess if the digital images are acceptable (for example no obvious error in the digital images). " in addition, EMI will have to bear the cost to ensure surveillance and maintenance of the cameras, to keep them clean of any obstructions and to check that they have not been misaligned or damaged by adverse weather conditions. EPA predicted uWe also believe that more state and local rules will require DCOT for the determination of opacity, which will provide incentive for competitors to enter the market place" response to comments at 14. it is worth outlining that in the 2 years since the Ferroalloys NESHAP rules were published, no competitor to Virtual Technology lias emerged that would allow EMI to negotiate prices for DCOT implementation and operation. D, Other Agency Contacts Edward Nam Director, Air and Radiation Division US EPA Region 5 77 W. Jackson Blvd. Chicago, JL 60604 Robert Hodanbosi Chief, Division of Ait Pollution Control Ohio Environmental Protection Agency Ohio EPA - DAPC Lazarus Government Center 50 W. Town Si, Suite 700 P.O. Box 1049 Columbus, OH 43216 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00009 The opacity standard in the Ferroalloys NESHAP was developed with the express expectation that it would be enforced via Method 9, just as comparable emission standards have been and are enforced at facilities throughout the United States. In specifying DCOT in the final rale, EPA did not identify any unique characteristics of fugitive particulate emissions from ferroalloys production that rendered Method 9 unsuitable or unreliable for enforcing the rule's opacity standard. Indeed, to certify DCOT for use at EMIT facility in accordance with ASTM standards, EMI would need to demonstrate that DCOT accurately replicated Method 9 measurements. As EPA continues to recognize as valid, reliable test method for enforcement of opacity standards like the standard in the Ferroalloys NESHAP, Method 9 clearly satisfies the criteria set forth in EMC GD-022R4 for approval as an alternative test method. Respectfully submitted, a'Y-fiItf fmure Guillot Chief Executive Officer Eraniet Marietta Inc. J 6705 State Route 7 Marietta, OH 45750 (740) 376-5914 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00008006-00010