


DATE:	January 5, 2017

SUBJECT:	Development of the RTR Proposal Risk Modeling Dataset for the Friction Materials Manufacturing Facilities Source Category

FROM:	David Bullock, RTI International
	Cassy Becker, RTI International

TO:	Susan Fairchild, Environmental Protection Agency
	Korbin Smith, Environmental Protection Agency

I.	BACKGROUND
Under section 112(f)(2) of the Clean Air Act (CAA), the U.S. Environmental Protection Agency (EPA) is directed to conduct risk assessments on each source category subject to maximum achievable control technology (MACT) standards and determine if additional or more stringent standards beyond those originally promulgated are needed to reduce any residual risk remaining after promulgation. The EPA is directed to conduct the section 112(f)(2) residual risk review eight years after promulgation of the original standards. In addition, section 112(d)(6) of the CAA requires EPA to review and revise the MACT standards, as necessary, taking into account developments in practices, processes, and control technologies. The EPA promulgated the national emissions standards for hazardous air pollutants (NESHAP) for Friction Materials Manufacturing Facilities (40 CFR part 63, subpart QQQQQ) in the Federal Register on October 18, 2002. The purpose of this memorandum is to document development of the section 112(f)(2) residual risk modeling input file for subpart QQQQQ. 
II.	FACILITIES, EMISSION PROCESS GROUPS, AND CATEGORY-SPECIFIC EMISSIONS
Facilities Modeled
Facilities that manufacture friction materials using a solvent-based process and that are major sources of HAP are subject to the Friction Materials Manufacturing NESHAP (63QQQQQ).  Only two facilities have been identified with processes subject to 63QQQQQ, as shown in Table 1. Risk assessment inputs were developed for both of these facilities. The search for major source friction materials manufacturing facilities with solvent-based processes subject to 63QQQQQ is documented in a separate memorandum.[1] No area sources or synthetic area sources were included in the risk modeling because none were identified with solvent-based processes subject to emissions standards under 63QQQQQ.
Emission Process Groups Modeled
The Friction Materials Manufacturing NESHAP applies to each new, reconstructed, or existing solvent mixer located at a friction materials manufacturing facility. Separate emission limitations are established for large solvent mixers and for small solvent mixers. Both of the identified facilities subject to 63QQQQQ have large solvent mixers; neither of the identified facilities has a small solvent mixer. As such, the only emission process groups modeled are large solvent mixers. The EPA does not believe there to be other significant sources of HAP at friction materials manufacturing facilities. Table 2 presents a summary of the friction materials manufacturing emission process groups and the number of emission points that were modeled for each facility.
Category-Specific Emissions for Subpart QQQQQ Sources
The pollutants regulated under subpart QQQQQ are HAP solvents. A HAP solvent is defined in 63QQQQQ as, "a solvent that contains 10 percent or more of any one HAP, as listed in section 112(b) of the Clean Air Act, or any combination of HAP that is added to a solvent mixer."[2] Emissions standards for the affected friction materials manufacturing sources are presented in Table 3.
The only HAP solvent emitted from affected sources under 63QQQQQ at Railroad Friction Products Corporation (RFPC) is hexane.  No other HAP emissions were reported from the affected sources.
The only HAP solvent emitted from affected sources under 63QQQQQ at Knowlton Technologies LLC (Knowlton) is methanol.  In addition to HAP solvent emissions, other HAP's are emitted from resins used within the source category including phenol and formaldehyde.
Emissions estimates were developed for hexane, methanol, phenol, and formaldehyde, as appropriate to each facility.
III.	EMISSIONS DATA COLLECTION
Because there are only two facilities currently subject to 63QQQQQ, EPA elected to conduct site visits to each of the two facilities, in lieu of developing a section 114 information collection request for the source category. Each site visit request letter included an Enclosure 1 with site-specific questions designed to gather the necessary information to support the risk and technology review.[3,4] Site visit reports were developed to document the information collected during the visits and provided in supplemental requests.[5,6]
Examples of the types of data collected relevant to the risk modeling inputs include the following:
Types and number of equipment (emission units) subject to 63QQQQQ;
Locations of emission points;
Reported actual annual emissions of HAP released from affected sources under 63QQQQQ, and emission calculation methodologies;
Emission control measures for each emission point;
Stack height, exhaust gas temperature, stack diameter, exit gas velocity, and exit gas flow rate for stack releases; and
Operating schedule (hours per day, days per week, weeks per year, hours per year).
These facility-specific data were reviewed then incorporated into the risk modeling file, as appropriate. 
 IV.	EMISSION POINT LOCATIONS, STACK PARAMETERS, AND DEFAULT VALUES
 Emission Point Locations
Both modeled facilities provided engineering drawings and aerial photos indicating the locations of relevant emission points. Emission point locations were plotted using Google Earth(TM) and sent to each facility for confirmation or correction. The confirmation or correction consisted of sending individual facilities a picture (.jpg) file of the locations plotted in Google Earth(TM), with a request that the facility either confirm or correct the plotted locations. In addition, the results plotted in Google Earth(TM) were checked to ensure that the emission point latitudes and longitudes were located on a building at the facility site, or at a control device near a building. Where the results were not located on or near a building at the facility site or a control device, the latitudes and longitudes were adjusted.
 Stack Parameters
Stack parameters necessary as modeling inputs include stack height, exit gas temperature, stack diameter, exit gas velocity, and exit gas flow rate. Actual values for these parameters for each relevant emission point were collected during the site visits to the two modeled facilities.  Exit gas flow rates were calculated from the stack diameter and exit gas velocity.  Modeled stack parameters for the relevant emission points at the two facilities are presented in Table 4.
Default Values
Actual stack parameters were obtained for all modeled emission points, with one exception. RFPC did not have an actual stack diameter or exit gas velocity for their hexane storage tank (ES-50). The "Default Stack Parameters SCC" table found in the EPA's "Stack Parameter Default 05102015" database was reviewed to select default values for these two parameters.[7] This emission point is most similar to Source Classification Code (SCC) 40701616 which represents "Petroleum and Solvent Evaporation; Organic Chemical Storage; Fixed Rood Tanks  -  Alkanes (Paraffins); Hexane: Working Loss." The default stack diameter is 0.3 feet. The default exit gas velocity is 0.389 feet/second. These values were incorporated into the modeling file for RFPC. The exhaust gas flow rate for ES-50 was calculated from the default stack diameter and exit gas velocity.
 V.	DEVELOPMENT OF EMISSION RATES FOR MODELING
 Actual Emissions
Actual emission rates developed for this data set were based on each facility's reported actual calendar year (CY) 2015 emissions.
RFPC
For RFPC, actual annual emissions of hexane are estimated based on a mass balance approach, consistent with the requirements of §63.9520. On a daily basis RFPC measures and documents the daily pounds of hexane introduced into the solvent mixer, the daily pounds of solvent recovered by the condenser system, the daily pounds of non-recovered hexane (hexane introduced, minus hexane recovered), and the average hexane recovery percentage. In CY2015, reported facility-wide emissions of hexane were 466,078 pounds (233 tons).
The facility-wide emissions of hexane are then apportioned among a series of emission points throughout the facility based on the results of an engineering study, as documented in their annual emissions calculations. A summary of the calculation of apportioned RFPC CY2015 actual emissions is presented in Table 5.
Knowlton
For Knowlton, actual annual emissions of formaldehyde methanol, and phenol are estimated based on the total annual purchase quantity of resin/solvent containing HAP, the maximum HAP contents percentage for each of the resin/solvents, the permanent total enclosure capture efficiency (100 percent), and pollutant-specific incinerator/boiler control efficiencies, in accordance with their Alternative Control Plan, consistent with the requirements of §63.9570. The saturator process line, including the resin kitchen, is located in a permanent total enclosure, therefore the assumed capture efficiency for emissions is 100 percent. Emissions from the permanent total enclosure are routed through an incinerator/boiler. Emissions testing conducted in January 2003 was used to determine the incinerator/boiler destruction efficiencies for a select group of organic compounds, including formaldehyde, methanol, and phenol, applied in the CY2015 air emissions inventory. Table 6 presents a summary of the calculation of CY2015 actual emissions for Knowlton.
Allowable Emissions
Allowable emissions for this data set were based on either federally-enforceable Title V permit emission limits, or on the applicable emission limits under §63.9500.
RFPC
      Allowable emissions for CY2015 were set equal to actual emissions. This was done due to the facility utilizing solvent substitution to comply with the MACT standard. Solvent substitution credits the facility for 100 percent recovery on every batch not using HAP solvent. This 100 percent recovery is then averaged with the actual percent recovery of the solvent condenser used in HAP solvent batches to determine the facilities total emission percentage. The calculated average does not properly represent the true emissions of the industry, resulting in our decision to set actual emissions equal to allowable emissions. 

Knowlton
Allowable emissions for CY2015 were estimated by scaling the actual emissions rates, assuming that Knowlton achieved only the minimum required recovery fraction (70 percent) under §63.9500, rather than the actual capture and control efficiencies (>96 percent) achieved for each pollutant.  Table 7 presents a summary of the calculation of CY2015 allowable emissions for Knowlton.
Acute Emissions
Emissions from equipment at friction materials manufacturing facilities can vary over time. In order to evaluate the acute effects of these variations, annual actual emission values were multiplied by the acute RTR default factor of ten.
To develop screening estimates of acute exposures in the absence of hourly emissions data, the EPA generally develops estimates of maximum hourly emissions rates by multiplying the average actual annual hourly emissions rates by a default factor to cover routinely variable emissions. The EPA chooses the factor based on process knowledge and engineering judgment reflecting, where appropriate, circumstances of the particular source category.
The default factor reflects the findings of a Texas study of short-term emissions variability, which showed that most peak emission events in a heavily-industrialized four-county area (Harris, Galveston, Chambers and Brazoria Counties, Texas) were less than twice the annual average hourly emissions rate.[8] The highest peak emissions event was 74 times the annual average hourly emissions rate and the 99th percentile ratio of peak hourly emissions rate to the annual average hourly emissions rate was 9. Considering this analysis, to account for more than 99 percent of the peak hourly emissions, we apply a conservative screening multiplication factor of 10 to the average annual hourly emissions rate in our acute exposure screening assessments as our default approach. However, we use a factor other than 10 if we have information that indicates that a different factor is appropriate for a particular source category. For this source category, there was no such information available and the default factor of 10 was used in the acute screening process. 
VI.	QUALITY CONTROL CHECKS AND RESULTING CHANGES
Data reported and incorporated into the risk modeling file were checked in several ways. Data quality checks that were performed and changes that were made are briefly summarized below:
Source latitudes and longitudes reported by facilities were checked in Google Earth(TM) to verify accuracy and were corrected as needed. Correct format for latitude and longitude was reviewed and corrected as needed.
The total number of emission points were checked in Google Earth(TM) and with facility contacts to verify accuracy and were corrected as needed.
Reviewed, identified and corrected missing stack parameters (diameter, flow rate, temperature).
Reviewed and corrected pollutant names and pollutant codes.
Reviewed and corrected exhaust temperatures, stack heights, stack diameters, and exhaust flow rates that appeared to be out of standard ranges.
Verified that actual emissions were less than or equal to allowable and acute emissions.
Verified that multiple emission points of a given type (i.e., press line vents, post bake oven stacks) have identical emission rates for a given pollutant.
         Reviewed calculation methodologies and verified that methodologies were sound and that calculated emission rates were accurate.
The quality control efforts employed for this project resulted in a more accurate CY2015 data set. 
VII.	REFERENCES
Memorandum, David Bullock and Cassy Becker (RTI International) to Susan Fairchild and Korbin Smith (U.S. EPA), February 1, 2017. Subject: Identification of Major Sources for the NESHAP for Friction Materials Manufacturing (40 CFR Part 63, Subpart QQQQQ). Docket: EPA-HQ-OAR-2017-0358 
National Emission Standards for Hazardous Air Pollutants for Friction Materials Manufacturing Facilities. See Federal Register Notice 67FR64506, dated October 18, 2002.
Letter, Keith Barnett (U.S. EPA) to Greg Thompson (RFPC), "Site Visit Letter Railroad Friction Products Corporation," August 3, 2016. Docket: EPA-HQ-OAR-2017-0358
Letter, Keith Barnett (U.S. EPA) to Erick Austin (Knowlton), "Site Visit Letter Knowlton Technologies LLC," October 4, 2016. Docket: EPA-HQ-OAR-2017-0358 
Site visit report, Railroad Friction Products Corporation, January 31, 2017. Docket: EPA-HQ-OAR-2017-0358
Site visit report, Knowlton Technologies, LLC, January 31, 2017. Docket: EPA-HQ-OAR-2017-0358
Stack Parameter Default 05102015.accdb - Default Stack Parameters SCC table.
Variable Industrial VOC Emissions and Their Impact on Ozone Formation in the Houston Galveston Area, Texas Environmental Research Consortium, April 16, 2004. Docket: EPA-HQ-OAR-2017-0358


                                        
                                        
 Table 1.	List of Friction Materials Manufacturing Facilities Modeled for Residual Risk
                                      RTI
                                    Number
                                  NEI Site ID
                             EIS Facility Site ID
                             Facility Registry ID
                                 Facility Name
                                Facility Address
                                 Facility City
                                Facility State
                               Facility Zip Code
                           Major/Area Source Status
                                       1
NEI43339
                                    8376611
                                 110000349980
 Railroad Friction Products Corp
 13601 Airport Road
 Maxton
                                      NC
                                     28352
                                     Major
                                       2
NEI36019
                                    8036511
                                 110000325988
 Knowlton Technologies LLC
 213 Factory St
 Watertown
                                      NY
                                     13601
                                     Major

                                       
                                       
Table 2.	Summary of Friction Materials Manufacturing Emission Process Groups and Emission Points Modeled
                                      RTI
                                    Number
                                  NEI Site ID
                                 Facility Name
                      Emission Process Group Description
                         Emission Process Group Count
                            No. of Emission Points
                                       1
NEI43339
Railroad Friction Products Corp
Large Solvent Mixer
                                       1
                                      21
                                       


Small Solvent Mixer
                                       0
                                       0
                                       2
NEI36019
Knowlton Technologies LLC
Large Solvent Mixer
                                       2
                                       1
                                       


Small Solvent Mixer
                                       0
                                       0

Table 3.	Emission Standards for New and Existing Affected Sources under Subpart QQQQQ [a]
 Affected Source/ Emission Unit
                                   Pollutant
                                Emission Limit
                                Averaging Time
 New, reconstructed, or existing large solvent mixer
                                  HAP Solvent
 Limit HAP solvent emissions to the atmosphere to no more than 30 percent of that which would otherwise be emitted in the absence of solvent recovery and/or solvent substitution
 7-day block average
 New, reconstructed, or existing small solvent mixer
                                  HAP Solvent
 Limit HAP solvent emissions to the atmosphere to no more than 15 percent of that which would otherwise be emitted in the absence of solvent recovery and/or solvent substitution
 7-day block average
  [a] §63.9500(a) and (b).
  

                    Table 4.	Summary of Stack Parameters for Modeled Facilities
                               Emission Unit ID
                  Affected Source/ Emission Unit Description
                                   Longitude
                                (X coordinate)
                                   Latitude
                                (Y coordinate)
                                 Stack Height,
                                      ft
                            Exit Gas Temperature, F
                              Stack Diameter, ft
                           Exit Gas Velocity, ft/sec
                        Exhaust Gas Flow Rate, ft3/sec
Railroad Friction Products Corporation
ES-10
 Sigma mixer condenser vent
-79.367764
 34.762301
49.5
54
0.33
 10.9
 0.93
ES-03
 Wet mix room exhaust vent
-79.367654
 34.762412
48.5
100
5
 20.3
 398.59
ES-26
 Press line 2
-79.367794
 34.762029
32
100
1.5
 20.8
 36.76
ES-27
 Press line 3
-79.367782
 34.762092
32
100
1.5
 20.8
 36.76
ES-28
 Press line 4
-79.367770
 34.762149
32
100
1.5
 20.8
 36.76
ES-29
 Press line 5
-79.367761
 34.762207
32
100
1.5
 20.8
 36.76
ES-31
 Press line 7
-79.367943
 34.762109
32
100
1.5
 20.8
 36.76
ES-32
 Press line 8
-79.367978
 34.762196
32
100
1.5
 20.8
 36.76
ES-33
 Press line 9
-79.368001
 34.762241
32
100
1.5
 20.8
 36.76
ES-25
 Press line 1
-79.367806
 34.761961
32
100
1.5
 20.8
 36.76
ES-30
 Press line 6
-79.367922
 34.761946
32
100
1.5
 20.8
 36.76
ES-34
 Press line 10
-79.367981
 34.761899
32
100
1.5
 20.8
 36.76
ES-35
 Press line 11
-79.367820
 34.761902
32
100
1.5
 20.8
 36.76
ES-20
 Post bake oven 1
-79.367783
 34.761798
32
200
0.8
 10
 5.03
ES-21
 Post bake oven 2
-79.367832
 34.761803
32
200
0.8
 10
 5.03
ES-22
 Post bake oven 3
-79.367876
 34.761807
32
200
0.8
 10
 5.03
ES-23
 Post bake oven 4
-79.368082
 34.761830
32
200
0.8
 10
 5.03
ES-24
 Post bake oven 5
-79.367988
 34.761817
32
200
0.8
 10
 5.03
ES-51
 NABCO oven
-79.367952
 34.762083
32
200
0.8
 10
 5.03
ES-50
 Hexane storage tank
-79.367532
 34.762701
13.8
Ambient
0.3
 0.389
 0.03
ES-61
 Sigma mixer grinder
-79.367452
 34.762284
32.5
Ambient
1
 49.2
 38.64
Knowlton Technologies, LLC
1-SVSAT
 Saturator Line
-75.906537
 43.976514
95
465
5.2
 36.8
 781.53
 
        Table 5.  Summary of Calculation of RFPC Actual Emission Rates

                                       


       Table 6.	Summary of Calculation of Knowlton Actual Emission Rates
                                       

     Table 7.	Summary of Calculation of Knowlton Allowable Emission Rates

                                   Compound
                             Actual Emissions, TPY
                   Actual Capture and Control Efficiency, %
                          Uncontrolled Emissions, TPY
                  Allowable Capture and Control Efficiency, %
                           Allowable Emissions, TPY
Formaldehyde
                                     0.49
                                     96.6
                                     14.41
                                      70
                                     4.32
Methanol
                                     6.27
                                     99.3
                                    895.71
                                      70
                                    268.71
Phenol
                                     0.44
                                     99.8
                                    220.00
                                      70
                                     66.00


