MEMORANDUM

TO: 		EPA Docket No. EPA-HQ-OAR-2020-0505

FROM: 	Korbin Smith, Physical Scientist
      U.S. EPA/OAQPS/SPPD/MMG

DATE: 	October 13, 2020

SUBJECT: Development of the RTR Proposal Risk Modeling Dataset for the Carbon Black Production Source Category.	

												

1.0 PURPOSE
      This memorandum presents our analyses related to both the development of the MACT-allowable emissions levels used to assess risk as well as the derivation of the hourly emissions rates used to estimate acute risk from the carbon black manufacturing source category. These analyses are being done in order to inform and satisfy EPA's statutory obligations under section 112(f)(2) of the Clean Air Act.
      
2.0 FACILITIES, EMISSION PROCESS GROUPS, AND CATEGORY-SPECIFIC EMISSIONS

A.	Facilities Modeled

      Carbon black production facilities included as part of the source category are facilities that produce carbon black via any of the three processes: furnace black, thermal black, or acetylene decomposition. Upon review, 15 identified carbon black production major source facilities are owned by 5 companies across 6 states, with facilities being primarily located in Texas and Louisiana, but also located in Kansas, Alabama, Oklahoma, and Ohio For more information on the major sources of carbon black production please see the memo, "Carbon Black Major Source Memo," which is available in the docket for this rulemaking.
B. 	Emission Process Groups 
      Emissions data was received from the 2016 National Emissions Inventory (NEI). Emissions were categorized into process groups by using source classification codes (SCCs) and unit descriptions. Process groups include combustion device, process filter, fugitive emissions, purge filter, and storage tank.  The process groups are used to understand the relative contribution of each type of process to the overall risk results after modeling is complete. 
C.	Speciation Profile
      Mercury emissions were speciated using the profile 50/40/10 for Elemental, Gaseous, Particulate Mercury. 
2.0 MACT-ALLOWABLE EMISSIONS ESTIMATES
      Available emissions data that form the basis of the inputs to the risk modeling file dataset are estimates of the mass of hazardous air pollutants (HAP) emitted during a specified annual time period. These emissions are commonly referred to as actual emissions, and in some cases these actual emissions levels can be less than the emissions limitations in a maximum achievable control technology (MACT) standard. The emissions level allowed to be emitted by a MACT standard (e.g., the MACT "floor") is more commonly referred to as the MACT-allowable emissions level for purposes of residual risk review.
      The basic approach used to estimate MACT-Allowable emissions for the carbon black source category was to, first, start with the available emissions data estimates representing actual emissions and assess whether these data were reasonable estimates of the levels allowed by the MACT standards for emissions sources regulated by the national emission standards for hazardous for air pollutants (NESHAP) for carbon black manufacturing (i.e., 40 CFR part 63, subpart YY). Table 1 summarizes the formats of the MACT standards for the emissions sources addressed in the NESHAP for carbon black. 
      
       Table 1. Formats of MACT Standards for the NESHAP for Ethylene Production

 Emissions Source
 Format of Primary Standard
 Main Unit Filter Process Vents
 Performance Standard or Concentration Limit
 Equipment Leaks
 Work Practice Standards
 Other Process Vents
 N/A

      The ability to estimate MACT-allowable emissions from the actual emissions dataset is largely dependent on the format of the standard for a given emissions source as well as types of controls employed by the source. With respect to the various types of controls used within the carbon black source category, none is more prevalent than the use of a combination of combustion control devices (CCD's). Facilities that manufacture carbon black typically have several types of CCD's including but not limited to, flares, incinerators, boilers, and dryers. The CCD's can be used to control emissions for a single emissions source, or as is generally the case, to control emissions from multiple emission sources/emission source types.
      CCD's are designed to handle a large range of variable flowrates and compositions of combustible waste gases and, within the carbon black source category, generally control emissions from multiple emission source types. Consideration of this, along with not having a specific limit on how much gas can be combusted in a CCD (given that in many cases multiple emissions sources are being controlled by this control device), makes it extremely difficult to determine an allowable emission rate. 
      Historically, the majority of facilities in the carbon black source category utilized flares to control emissions.  Emissions inventories for flares in the carbon black source category are developed using engineering knowledge and, in many instances, presume this 98 percent level of control (e.g., see as an example Technical Supplement 4: Flares in "2016 Emissions Inventory Guidelines," (TCEQ. 2017)), and the agency is unaware of any data that suggests that CCD's in the carbon black source category are consistently over-controlling HAP emissions beyond 98 percent.
      For purposes of the MACT-allowable risk analysis, we evaluate whether it is necessary to tighten the existing MACT standard. Thus, weighing all of these factors for combustion devices, we believe that the actual emission levels are a reasonable estimation of the MACT-allowable emissions levels where the performance standards allow the use of a flare or to reduce emissions of total HAP by 98 weight-percent. 
      For equipment leaks, which are subject to work practice standards, there would be no difference between actual and MACT-allowable emissions for facilities in the carbon black source category, provided the facilities are in compliance with the NESHAP as well as not conducting additional work practices proven to reduce emissions beyond those required in the rule. Therefore, weighing these factors for equipment leaks, we believe that the actual emission levels are a reasonable estimation of the MACT-allowable emissions levels allowed by the standard for the source category.
	
3.0 ACUTE RISK EMISSIONS ESTIMATES
      To develop estimates of acute exposures in risk and technology review (RTR) rulemakings, the agency generally assumes the 1-hr emissions rate for any emission point could be 10 times higher than its average hourly emissions (calculated by dividing the actual emissions by 8760 hours per year) in situations where the EPA lacks sufficient information on hourly emissions for given emissions sources. The basis for this assumption was derived from an analysis of short term release information collected from a Texas study of facilities in a four-county area (Harris, Galveston, Chambers, and Brazoria Counties, Texas) which were then compared against routine emissions rates for an entire facility. The conclusions for this analysis were that hourly emissions from any single release event to the average annual volatile organic compound (VOC) release rate for an entire facility were seldom greater than a factor of 10. 

A.	Approach to Derivation of Hourly Emissions Rates
      
      For the carbon black source category, a number of specific factors can be considered with respect to certain emissions sources in the source category to help fine tune and derive a conservative hourly emissions rate multiplier to assess acute risk rather than use a default factor of 10. These specific factors include, but are not limited to: process knowledge, the modes of operation (e.g., continuous, batch, intermittent, etc.) of certain unit operations and their connection to emissions, and other specific factors that would affect variability in hourly emissions from a given emissions source (e.g., meteorological conditions for storage vessels). Thus, in some cases, a factor of 10 may overestimate emissions while in others it may underestimate emissions for the upper range of expected variability for certain emissions sources. As such, consideration of these relevant factors, if known, should be applied to more accurately assess acute risk. 

3.1 Acute Risk Factors for the Carbon Black Source Category

Based on the approach described in section 3.1, we determined that the acute risk multipliers identified in Table 2 should be used in lieu of assigning a default factor of 10 to estimate hourly emissions and assess acute risks for each emissions source in the carbon black source category. The reasoning for changing the acute multiplier from the default (if it was changed) is also described in Table 2.
Table 2 - Summary of Acute Risk Multipliers Developed to Assess Acute Risks for the Carbon Black Source Category
Emissions Source
Acute Multiplier
Reasoning for Acute Multiplier
Main Unit Filter Process Vents and Other Process Vents (Furnace Black)
2
For process vents using the furnace black production process, these vent streams would be associated with continuous operations and essentially be steady state. Typical source variability would be minimal given their continuous nature. It us Thus, an hourly multiplier of 2 is applied for this emissions source.
Equipment Leaks
2
For fugitives from equipment leaks, we note that the current methods of estimating emissions make use of correlation equations in conjunction with EPA Method 21 readings that ultimately provide an hourly emission rate for the monitoring period in question. The actual emission rates estimated based on the direct EPA Method 21 readings are commonly divided by two to estimate the average emission rate between monitoring intervals (i.e., assuming the leak started mid-way between monitoring intervals as described in the emissions protocol). We expect the emission estimates from the direct EPA Method 21 readings to provide a direct measure of the maximum hourly emissions (as leak repair will be applied to reduce hourly emissions immediately after a monitoring cycle). Therefore, we applied an hourly multiplier of 2 to estimate maximum hourly rates from the annual average emission rates for this emissions source.

