          Responses to Comments - SO2 Data Requirement Rule Modeling
              (Source: Emails between Rick Gillam, EPA Region 4, 
                     and Tom Anderson, NCDENR, June 2017)
                                       
Duke Energy Marshall Steam Station
 
1)      Marshall Source Parameters
-          *Actual SO2 emissions were used in the impact assessment.  Information on the quality and completeness of these data should be provided as well as the procedures used to replace/substitute missing actual emissions.
SO2 emissions were obtained from Duke Energy staff.  
Duke Response: The SO2 emissions data were obtained from the data reported by our CEMS, using the same data base that is reported to the EPA Clean Air Markets Division (CAMD).  All substituted values are as reported to CAMD in accordance with Part 75 requirements.  SO2 data availability was 98.9% or greater for all stacks during this period.
-          *Varying stack exit parameters were indicated used in the modeling.  Information on the quality and completeness of these data should be provided as well as the procedures used to replace/substitute missing actual values.
Hourly varying stack parameters were obtained from Duke Energy staff.  
Duke Response:  The hourly stack exit velocity was calculated using Part 75 CEMS data.  CEMS monitor data availability for flow measurement was 99.5% or greater for all stacks during this period.  The hourly exit velocity was calculated from the hourly flow and stack temperature data  using the following equation. 
 
      V msec=A STD ft3hr x hr60 min x min60 sec x B+480528 x PI x D22x 0.3048 m3ft3  
       
      Where:
      V is the stack exist velocity in meter/sec
      A is the Part 75 stack flow rate in standard cubic feet per hour
      B is the stack gas temperature in degree F
      D is the stack diameter in feet
 
 
2)      Other Nearby SO2 Emission Sources
-          *The emission inventory of all nearby emissions sources considered should be provided.  The sources eliminated should be identified along with the basis for their elimination.  This information is needed to evaluate the appropriateness of the procedures used.  
A file including all SO2 sources within 50 km is attached to this message (filename: Duke SO2 Inventories_Actuals.xls).  Section 9, page 8, of the modeling report describes how nearby sources were evaluated for inclusion in the analysis.
 
Duke Energy Allen Steam Station
 
1)      Other Nearby SO2 Emission Sources
-          *The emission inventory of all nearby emissions sources considered should be provided.  The sources eliminated should be identified along with the basis for their elimination.  This information is needed to evaluate the appropriateness of the procedures used. 
A file including all SO2 sources within 50 km is attached to this message (filename: Duke SO2 Inventories_Actuals.xls).  Section 9, page 9, of the modeling report describes how nearby sources were evaluated for inclusion in the analysis.
 
    
Duke Energy Mayo Electric Generating Plant
    
1)      Modeled Emissions
-          *The Mayo facility was indicated to operate two coal-fired boilers each vented to individual stacks.  Figure 3 is indicated to provide the location of buildings, stacks, and the facility fence line.  Based on Figures 3 & 4 and the modeling files, there is only one stack.  These differences should be explained.
The text in Section 7, page 4, of the modeling report that indicates that the two boilers vent to individual stacks is an error.  The statement was simply an inadvertent carryover from another report that was used as a template and the two boilers do indeed vent from only a single stack.
 
2)      Other Nearby SO2 Emission Sources
-          *The basis for the hourly emission rates developed for the Duke-Roxboro Units 1-4 and CPI-Roxboro Units 1A, 1B, and 1C should be provided/ explained (e.g., actual emission from CEM data for period 2013-2015, etc.).  Additionally, the hourly emissions files used in the modeling should be provided for review.
A file including all SO2 sources within 50 km is attached to this message (filename: Duke SO2 Inventories_Actuals.xls).  Section 9, page 6, of the modeling report describes how nearby sources were evaluated for inclusion in the analysis.
 
The hourly emission rates for both the Duke  -  Roxboro units and the CPI units were obtained from actual CEM data for the period 2013-2015 and were obtained from EPA's Clean Air Markets Database (https://ampd.epa.gov/ampd/).  
 
The hourly emissions file used in the modeling is attached to this message (filename: mayo_hourly emissions SO2.txt)
 
 
Duke Energy Belew Creek Generating Station
 
1)      Modeled Emissions
-          * Missing CEMS stack temperatures were replaced with an average stack temperature determined from the CEMS data (i.e., 120 degrees F). If the data record has many missing values, use of the recorded actual stack temperatures about the missing value(s) to developed replacements would be more appropriate.
Duke Response:  Use of "before and after" temperatures might be appropriate for short periods of missing data where the operating conditions are similar during the missing period, but for longer periods of missing data, an average temperature is a more appropriate representation since the operating conditions (minimum to maximum generating load and heat input) may vary significantly over time .  Also, due to the continuous operation of the wet scrubbers, there is a relatively small range of temperature variation ( average 121 °F with standard deviation of 3.5 °F) and there should be little impact on the modeling whether using the average temperature or a temperature based on the period before and after. 
 
-          *These Auxiliary Boilers were modeled using the EPA Clarifications Memorandum guidance on intermittent sources that annualizes the emissions for the full year, to obtain an annual average hourly value used for each hour of the year.  These boilers were indicated to only operate when the coal-fired Utility Boilers are not operating or just starting up (i.e., no simultaneous operations).  If simultaneous operations are not possible and the coal-fired boilers have much larger SO2 hourly emissions than fuel oil-fired boilers, the operation of Auxiliary Boilers (ES-3 and ES-4) should not be controlling.  A comparison of the maximum SO2 emission rates for these boilers should be provided.  Because of the lower stack height for ES-3/4 boilers, the maximum impacts of the oil-fired and coal-fired boilers may be needed to demonstrate the appropriateness of treating the Auxiliary Boilers as intermittent sources. 
Duke Response:  The two auxiliary boilers combust fuel oil that is exclusively the ULSD fuel oil supply that is commercially available in the region. The maximum emissions rate for these boilers is 0.27 lb/hr each.  The emissions rate for the Belews Creek Units 1 and 2 EGU boilers operating at maximum averages approximately 1,200 lb/hr, with a typical range (95% of data) between 600 to 2000 lb/hr.
 
 
2)      Other Nearby SO2 Emission Sources
-          *The emission inventory of all nearby emissions sources considered should be provided.  The sources eliminated should be identified along with the reason/basis for their elimination to support the appropriateness of the procedures used. 
Duke Response:  A file including all SO2 sources within 50 km is attached to this message (filename: BCInventory010517.xlsx).  Section 13, pages 22-25, of the modeling report describes how nearby sources were evaluated for inclusion in the analysis.
 
-          *Documentation on the 2015 permit removal of the 4 coal-fired boilers at Millers Coors Brewery should be provided to support this source's exclusion from the modeling assessment..
Attached are two files to support the exclusion of the boilers from the modeling  -  1) a 2015 Permit Review in which removal of the boilers is discussed and, 2) the latest 2017 Compliance Inspection Report that shows all permitted sources at the facility.
-          *Section 13.1 indicates maximum annual emissions over the 2013-2015 period were used in the modeling of nearby emission sources.  This does not appear to agree with Section 13.0 indication that actual emissions were used.  The report should clearly indicate the basis of the emissions for each nearby source used in the screening process and in the impact modeling assessment. Also, the following information is needed:
         o   The reason the assumption of 8,760 hours of operation for years 2013 -2015 is appropriate.
 The reason permit estimated emission rates for the melt, casting, and arc furnaces emitted from roof vents are appropriate for permit application and not actuals during the 2013-2015 period. 
Duke Response: The modeled emissions rate was estimated by dividing the reported annual emissions over the period from 2013 to 2015  by 8760 hour/year. The inventory did not provide information on the hours of operation,  therefore  8,760 hours of operation was used to estimate the lbs/hr emissions rate used in the modeling . 
 
 



