5301 32nd Ave S
Grand Forks, ND 58201-3312
Phone 701.795.4000 
www.minnkota.com

5301 32nd Ave S
Grand Forks, ND 58201-3312
Phone 701.795.4000 
www.minnkota.com

                                 July 29, 2022

Melanie King
Environmental Protection Agency,
Office of Air Quality Planning and Standards			Sent via email: king.melanie@epa.gov
109 T.W. Alexander Drive
RTP, NC 27711

RE: 	Response of Minnkota Power Cooperative Clean Air Act Section 114 Request

Dear Ms. King,
	On May 12, 2022, EPA issued a Section 114 request (Section 114 request) to Minnkota Power Cooperative, Inc. (Minnkota) for information related to the agency's review of the Mercury and Air Toxics Rule (MATS) under Clean Air Act Section 112(d)(6) related to hazardous air pollutant (HAP) emissions.  EPA is specifically interested in mercury emission control technologies for lignite coal-fired units.  Minnkota appreciates EPA's approval of the extension we requested until July 29, 2022 to be fully responsive to the request.  
	The Section 114 request requires the completion of an information survey to collect information on mercury emissions, related unit operation, and control technologies.  Minnkota provides the attached survey spreadsheet with non-CBI information and non-CBI attachment A.  CBI-protected information has been submitted to EPA in accordance with the instructions in the Section 114 Request and 40 CFR Part 2, Subpart B (40 CFR § 2.201 et seq.).
	Minnkota also responds with additional information concerning our mercury control equipment and its operation history.  This information is intended to provide EPA with clarity and context for the responses.  
 Background
	Minnkota Power Cooperative is a not-for-profit electric generation and transmission cooperative headquartered in Grand Forks, North Dakota.  Minnkota provides wholesale electric energy to 11 member-owner distribution cooperatives located in eastern North Dakota and northwestern Minnesota.  Minnkota also serves as the operating agent for the Northern Municipal Power Agency (NMPA), headquartered in Thief River Falls, MN.  
	Minnkota is the operator and a partial owner of the Milton R. Young Station (the Young Station), a two-unit, cyclone lignite coal-fired power plant located near the town of Center, North Dakota.  Minnkota's electric generation portfolio also includes renewable energy, purchased primarily from three North Dakota wind farms, and hydroelectricity purchased from the Garrison Dam in central North Dakota.  In all, renewables comprise 42% of Minnkota's nameplate generation capacity.  Minnkota exists as a non-profit cooperative for the sole purpose of meeting the generation and transmission needs of the distribution cooperatives.
	Young Station Units 1 and 2 are well-controlled electric generating units, which provide energy to the Midcontinent Independent System Operator (MISO) system.  Young Station Units 1 and 2 have substantially reduced NOx and SO2 emissions, which have been documented in the context of the Regional Haze program.  Unit 1 has SO2 reductions of 96% since 2002, and Unit 2 has reduced SO2 75% since 2002.  Both Units reduced NOx emissions approximately 60% since 2002.  The Young Station has been in continuous compliance with MATS limitations since the effective date of the program. 

 Lignite Combustion at the Young Station
	Both boilers fire North Dakota lignite supplied from BNI Coal, Ltd.'s Center Mine.  The Young Station's lignite supply comes from a mine owned by BNI Coal Inc. in close proximity to the plant.  It is well-known and consistent with Minnkota's experience that lignite deposits vary in quality, including fuel combustion performance and mineral content.  Relevant to EPA's inquiry, mercury content in the lignite varies significantly.  Different areas within the mine yield lignite in varying quality content (including mercury) on a day-to-day basis.  Minnkota is adept at pivoting operations at Units 1 and 2 to accommodate for the changing lignite quality to assure compliance with MATS mercury limitations.  However, a compliance margin is critical.  
	The current MATS mercury requirements for lignite combusting units set an emissions limitation for low rank virgin coal that allows for variations in coal quality.  The current MATS mercury emission limitation provides lignite power plants enough leeway to account for high mercury batches of coal, recognizing that mercury emissions are higher from coal with lower heat values.  77 Fed. Reg. 9304, 9388 (Feb. 16, 2012).  Indeed, it is revealing that no lignite-burning facilities were in the top 12 performing percent of sources for mercury emissions when MATS was developed.  Id.  To address daily and weekly emissions fluctuations, a 30-day rolling averaging period is essential in addition to maintaining a higher mercury emission rate for this subcategory of fuel.  We ask that EPA consider the unique challenges inherent in combusting lignite.  
 Young Station Unit 1 and 2
	Unit 1 (257 MWg) commenced commercial operation in 1970, and Unit 2 (477 MWg) commenced operation in 1977.  Both units employ Babcock and Wilcox cyclone boilers, and both boilers fire North Dakota lignite supplied from BNI Coal, Ltd's Center Mine.  Each unit is equipped with Advanced Separated Over Fire Air (ASOFA) and Selective Non-Catalytic Reduction (SNCR) for NOx control.  Each unit is equipped with Wet Flue Gas Desulfurization (WFGD) for SO2 control.  Each unit employs an Electrostatic Precipitator (ESP) for particulate control.  Each unit employs halide injection and activated carbon injection for the control of mercury. 
	Currently, Minnkota uses the same mercury control strategies for both Units, as more fully described in Section IV.  
	Minnkota's Units operate differently.  Although Minnkota uses the same control devices for the Units, operation and emissions output differs based on a number of factors.  The Units vary in capacity and control device design.  Unit 2 has a different ductwork configuration between the air heater and the electrostatic precipitator than Unit 1.  Unit 1 has shorter ductwork.  Length of ducting effects the amount of residence time for the flue gas to be exposed to PAC, a longer duct provides more time for oxidized mercury to be adsorbed by the PAC. 
	EPA should acknowledge and consider that there is variability in Unit performance and emissions control capabilities from unit to unit.  In the Clean Air Act Section 112(d)(6) analysis, we ask EPA to consider unit differences.  Analysis of Unit 1 and Unit 2 demonstrate how seemingly similar unit (cyclone/lignite/Hg controls) perform differently.  The present MATS mercury emissions limitation allows for such unit variability.  
 History of Mercury Control and Monitoring Equipment at the Milton R Young Station
	Minnkota changed its equipment to control mercury since initial compliance with the MATS Rule.  The following section outlines these equipment changes.  We also highlight these changes to explain why some information in the Survey spreadsheet is not in Minnkota's possession.  
	A.	Mercury Control Equipment for Units 1 and 2
	Minnkota used the same mercury control equipment configurations for Unit 1 and Unit 2 during each of these time periods.
	Initial MATS Compliance (2015-2021)
	From July 24, 2013 through November 16, 2021, Minnkota purchased refined coal from GS RC Minnkota, LLC. Minnkota believes that GS RC Minnkota had certain rights to and operated, through Tinuum Group LLC (Tinuum), the fuel additive equipment located at the Young Station.  We believe that Tinuum operated the equipment to apply a fuel additive to the lignite called Cyclean B, which is a liquid halogen solution.  Additionally, Clean Coal Solutions, LLC (CCS) licensed to Minnkota a separate fuel additive system and supplied sufficient quantities of additional Cyclean B for application.  Minnkota operated this system, adding Cyclean B variably based on operating condition and coal quality.  Cyclean B is intended to reduce emissions of mercury.  Minnkota applied Cyclean B to the lignite prior to its arrival in the coal silo, in advance of lignite crushing before combustion in the boiler.
      On or around 2013, Minnkota procured a powdered activated carbon (PAC) system designed and installed by ADA-ES.  The equipment was placed in service, prior to the MATS initial compliance date.  Minnkota operated the PAC system in conjunction with the halogen skid described above.  The PAC system is currently in operation.  
      CCS, Tinuum and GS RC Minnkota, LLC are third-party companies not affiliated with Minnkota.  Since these companies installed and operated the refined coal system, Minnkota does not have the equipment design information, nor was there any performance specifications or guarantee.  Minnkota was never provided and does not have access to the refined coal system operating data and/or performance standards.  Based upon information and belief, Tinuum applied an average of 6.0 ppm of Cyclean B during the applicable one week period identified in the survey for Unit 1 and Unit 2; however, no reliable refined coal system operating data is in Minnkota's possession or otherwise available to Minnkota.
      In summary, from 2015 to November 2021, Minnkota reduced mercury emissions at Units 1 and 2 using a combination of: (1) Cyclean B applied by Tinuum; (2) Cyclean B applied by Minnkota; and (3) PAC applied by Minnkota.  
      MATS Compliance (November 2021 - Present)
      In November 2021, the equipment operated by Tinuum was decommissioned.  Minnkota transitioned mercury removal to a fuel additive system designed by NALCO.  The NALCO equipment was installed on or about March 2015 as a redundant system intended for MerControl 7895 injection; however, Minnkota did not operate the NALCO equipment for MATS compliance until November 2021. The system has never been used with MerControl 7895 for mercury reduction.  
      In November 2021, Minnkota altered the NALCO pumps for operation, and by June of 2022, Minnkota added additional flow meters and replaced the pumps to allow the equipment to operate efficiently with M-Prove.  Minnkota currently uses this system to apply M-Prove fuel additive procured from ADA.  Minnkota also continues to inject PAC post-combustion.  
      In summary, from November 2021 to present, Minnkota reduced mercury emissions at Units 1 and 2 using a combination of: (1) M-Prove applied by Minnkota; and (2) PAC applied by Minnkota.  Minnkota maintains design and operation data for the NALCO system, as modified.  This is Minnkota's current mercury reduction operating system.  
      B.	Mercury Monitoring Equipment
      From 2015 to November 2019, Minnkota utilized a mercury CEMS made by Tekran for MATS compliance.  In November 2019, Minnkota replaced the mercury Tekran CEMS with a sorbent trap system for measurement of mercury.  Each of these types of mercury CEMS measure mercury at the chimney. Minnkota does not monitor mercury concentrations in the lignite fuel. 
V.	Section 114 Spreadsheet Requests
      Minnkota provides this additional information to provide context for its responses to the Survey.  
      A. 	Mercury Control Info Tab
      Given the differences in design and independent operations in the two Young Station mercury removal systems, Minnkota provides two separate tabs for "mercury control info."  One tab is for the system installed and operating on Unit 1, and the second tab is for the system installed and operating for Unit 2. 
      We note that the original NALCO design was modified by Minnkota during the deployment period in November/December 2021.  For this reason, the system design entries for responses in lines 59 through 73 of the survey reflect the original design of the system and the November/December 2021 changes made to accommodate the M-Prove product.  This is the current system design. 
	Minnkota does not perform coal analysis or inlet testing for mercury.  Coal ash analysis lab reports available during the periods requested are attached Attachment 2.
	Minnkota selected March 1 through 31, 2022 for both Unit 1 and Unit 2 responses to line 32-43 of the respective spreadsheets, which includes data that has been certified and reported. Some of the factors considered by Minnkota in choosing the averaging period were selecting: (a) a 30-day average period rather than a shorter time period, (b) a time period representative of the current system of control in operation, and (c) a time period in which there was 100% availability of Unit 1 and Unit 2.  
	B.	Unit 1:  Week 1 Hourly Hg Emissions (6/13/19 through 6/19/19)
	The system operated by Tinuum described in Section IV.A was the mercury removal system in operation during this period.  Week 1 data from B1 fuel additive equipment are not representative of current operation of mercury control equipment because this system was decommissioned in November 2021 and replaced by the modified NALCO design.  
	Minnkota observes that this week period is not representative of an achievable mercury emissions rate, but rather is representative of the quality of fuel for that week.  Units were operating at or near normal-high loads during this period, PAC injection was at 0.0 lb/min and Minnkota's injection of halogen was 1.9 ppm. It is likely that the emissions during this 7-day period are lower than Minnkota's typical 30-day rolling average in large part due to low mercury levels in the coal, based on Minnkota's experience and our review of the PAC and halogen data. 
	C.	Unit 1:  Week 2 Hourly Hg Emissions (12/25/21 through 12/31/21)
	The re-purposed, modified NALCO system described in Section IV.A was the mercury removal system in operation during this period.  This is the current system in operation at Units 1 and 2.  The ADA chemical, M-Prove, is the chemical currently used in the fuel additive equipment by Minnkota. 
	Minnkota observes that this one-week period is not representative of an achievable mercury 30-day rolling emissions rate, but rather is representative of the quality of fuel for that week.  During this period, Minnkota was applying an average of .0114 lb/min of PAC and halogen at an application rate of 6 ppm. During this Week 2 period, Unit 1 operated at low load during part of the week, from 12/30-12/31 at an average generation rate of 110 MWg.  This lower load operation caused higher reported mercury emissions. Additionally, in light of Minnkota's operation of the mercury control system with both injection of PAC and halogen, it is likely that the coal combusted during this period had a higher concentration of mercury. 
	D.	Unit 2:  Week 1 Hourly Hg Emissions (2/18/21 through 2/24/21)
      The CCS system described in Section IV.A was the mercury removal system in operation during this period. Week 1 data from Unit 2 fuel additive equipment are not representative of current operation of mercury control equipment because this system was decommissioned in November 2021 and replaced by the modified NALCO design.  
      Minnkota observes that this one-week period is not representative of an achievable mercury emissions rate, but rather is representative of the quality of fuel for that week.  During this period, Minnkota was applying an average of 0.98 lb/min of PAC and halogen at an application rate of 1.9 ppm. 
      The Unit operated at normal-high load during the early part of Week 1, but dropped to low load during the latter part of this week 1 period, from 2/21 to 2/24 at an average generation rate of 380 MWg. However, despite a drop in CO2 flow (average value of 8.81%), Minnkota was able to maintain a relatively low average mercury emission rate of 2.17 lb/Tbtu. In light of Minnkota's experience operating the control system with both injection of PAC and halogen, it is likely that the coal combusted during this period had a lower concentration of mercury. 
      E. 	Unit 2:  Week 2 Hourly Hg Emissions (11/10/19 through 11/16/2019)
	The CCS system described in Section IV.A was the mercury removal system in operation during this period. Week 2 data from Unit 2 fuel additive equipment are not representative of the current operation of mercury control equipment because this system was decommissioned in November 2021 and replaced by the modified NALCO design.
      During this Week 2 period, the Unit operated at mostly normal-high load, at an average generation rate of 455 MWg. Minnkota observes that this one-week period is not representative of an achievable mercury emissions rate, but rather is representative of the quality of fuel for that week.  During this period, Minnkota was applying an average of 0.21 lb/min of PAC and halogen at an application rate of 1.9 ppm. In light of Minnkota's operation of the control system with both injection of PAC and halogen, it is likely that high mercury content in the lignite caused higher mercury emissions.
VI.	Important Observations regarding Data Provided In this Section 114 Response 
	Unit 1 and 2 mercury emissions and operational data for Week 1 and 2 are not representative of emissions achievable on a 30-day rolling basis for the Units, nor are the weeks comparable to one other. The significant differences between the two weeks are outlined below.  
 Most importantly, lignite coal quality varied significantly, including variability of mercury concentrations.  Coal quality between the two weeks differed based on our analysis of the sorbent and PAC applied as compared to the mercury emission rates, as discussed below.  
 Milton R. Young Unit 1 - Week 1 and Week 2 used different mercury control equipment and halogen manufacturers.  
 Milton R. Young Unit 2 - Week 1 and Week 2 used the CCS system, which is no longer operating and was replaced by the modified NALCO system in November 2021.  
 Unit operation varied.  Week 1 included low-load operation conditions, which were not present during Week 2.  Mercury emissions are reported as higher at lower loads.  
	In addition, a limited 7-day data set does not account for fuel quality variability.  While Minnkota does not test inlet mercury concentrations, we do have coal analysis of random samples taken from coal conveyor belt as the coal storage silos are filled. A maximum of six samples are collected daily and analyzed for proximate analysis. These samples do not include mercury concentrations but show considerable variation in ash content, Btu value, and other constituents.
	Minnkota further observes that units are unique, as demonstrated by data from Unit 1 versus Unit 2.  Minnkota utilizes the same mercury control strategies for the two Units.  However, Unit 2 consistently has higher mercury emissions in comparison to Unit 1. The differences in mercury emissions during the requested periods are impacted by operating characteristics of the Units, such as varying load levels/load swings, unit size, wet scrubber design differences, and ductwork configuration.  The inconsistent hourly emissions and mercury system data illustrate these differences and should assist EPA in understanding unit variability in general and with respect to Unit 1 and Unit 2 operations. For this reason, the responsive weekly data cannot reliably or reasonably support changes in permitted emission rates or permit conditions because this limited data set is not representative of current operations, fuel variability, and uncertainty in fuel quality.  These factors all contribute to differences in emission characteristics. 
	Finally, even though Week 1 and Week 2 are not comparable weeks, Minnkota reviewed the data regarding PAC used and sorbent used in its possession.  Week 1 reported lower mercury emissions in comparison to Week 2 values.  However, Minnkota did not inject any PAC during Week 1.  Minnkota injected 1.9 ppm of sorbent.  We believe that CCS did not inject more than 6 ppm of halogen, although no documentation of that quantity is available.  Lower mercury emissions were likely the result of a weekly coal batch that was lower in mercury content during Week 1.  In comparison, Minnkota injected PAC and halogen during Week 2.  Mercury emissions were higher, likely due to low load operation during that week and suspected coal quality variations.  
	Conversely, Unit 2 Week 1 PAC was injected at a higher rate than Week 2.  Week 1 had lower CO2 during most of this 7-day period and yet was able to maintain a mercury emission rate around 2.17 lb/Tbtu, while mercury emissions during Week 2 were around 3.0 lb/Tbtu with normal-high load operation and CO2 concentrations averaging 10%.  Minnkota attributes the inability to correlate the mercury emissions with the control system data to the following factors (1) variability of coal quality and (2) lack of information regarding Tinuum's operation of the halogen system during the pre-November 2021 operating period.  In other words, the lack of mercury control information for Unit 2 during Week 1 prevents us from drawing a conclusion regarding the variables that impacted mercury emissions that week.
VII.	Conclusion
	Thank you for your consideration of the information in this narrative and in the Survey.  If you have any questions, please contact Shannon Mikula at 701.795.4211 or smikula@minnkota.com regarding our responses.  


						Sincerely,
																	Craig J. Bleth
						Vice President of Power Supply
						Minnkota Power Cooperative, Inc.

CC: Shannon Mikula, smikula@minnkota.com
