
FROM:	Jeff Coburn  -  RTI International
TO:	Phil Mulrine, EPA/OAQPS
FOR:	EPA Docket No. EPA-HQ-OAR-2019-0373
DATE:	July 2, 2019
SUBJECT:	Control Cost Estimates for Organic HAP Emissions from Iron and Steel Foundries
1.	Purpose
      This memorandum documents the analysis conducted to assess potential additional organic emission control requirements as part of the ample margin of safety assessment for the national emission standards for hazardous air pollutants (NESHAP) for major source iron and steel foundries (40 CFR part 63 subpart EEEEE). 
2.	Background
      Section 112(d) (2) and (3) of the CAA directs the U.S. Environmental Protection Agency (EPA) to develop maximum available control technology (MACT) standards to control hazardous air pollutants (HAP) emissions from major sources. The term "major source" means any stationary source or group of stationary sources located within a contiguous area and under common control that emits or has the potential to emit considering controls, in the aggregate, 10 tons per year or more of any hazardous air pollutant or 25 tons per year or more of any combination of hazardous air pollutants. On April 22, 2004, the EPA published final standards for iron and steel foundries that are major sources of HAP emissions (40 CFR part 63, subpart EEEEE). 
      Section 112(f)(2) of the CAA requires the EPA to determine whether promulgation of additional standards is needed to provide an ample margin of safety to protect public health or to prevent an adverse environmental effect. Section 112(f)(2)(B) of the CAA further expressly preserves the EPA's use of the two-step approach for developing standards to address any residual risk and the Agency's interpretation of "ample margin of safety" developed in the Benzene NESHAP (54 FR 38044, September 14, 1989). In the ample margin of safety assessment, the EPA considers, among other things, additional standards that reduce the number of persons at risk levels higher than approximately 1-in-1 million, taking into account costs and economic impacts, technological feasibility, and other relevant factors.
      Based on the results of the residual risk analysis, several facilities, including the facility estimated to have the highest maximum individual risk (MIR), exhibited risks above 1-in-1 million due to the emissions of naphthalene and benzene. These organic HAP emissions occur from pouring, cooling and shakeout (PCS) lines and from mold- and core-making operations. 
3.	Control Options for Organic HAP Emissions from Iron and Steel Foundries
      Three potential emission reduction measures for organic HAP from PCS lines and mold- and core-making operations were identified: low-emitting binder systems, carbon adsorption, and incineration. Each of these options are discussed in further detail in the following subsections.
3.1	Low-emitting Binder Systems
      Low-emitting binder systems represent a pollution prevention method for reducing organic HAP emissions from mold- and core-making operations and PCS lines. Low-emitting binder systems may include inorganic binder systems or organic binder formulations with reduced levels of HAP and/or total organics. Reduced organic HAP content in the chemical binders leads to reductions in organic HAP emissions from the mold- and core-making operations. Organic HAP emissions from PCS lines are impacted by both the HAP content of the binders and the total organic content of the binders available for pyrolysis when exposed to molten metal. Therefore, a binder system with low HAP content but with a high overall organic content may still have substantial emissions during the PCS process. Thus, there are some difficulties determining whether an organic binder system is "low-emitting," and testing is generally needed to ensure an alternative organic binder system will reduce emissions for the facility when considering the overall process (mold- and core-making and PCS emissions combined). Inorganic binder systems, on the other hand, are generally effective at reducing HAP emissions from both mold- and core-making operations and PCS lines and may be considered "low-emitting" with limited or no additional testing. 
      Different binder systems exist because of their different properties and capabilities. The size, shape and tolerance of the castings, the production volume, and the environmental conditions (temperature and humidity) must all be considered when selecting a binder system. Some binder formulations may have poor performance when the humidity is high; some may be negatively impacted by high or low ambient temperatures; some may not have the strength needed for large castings, while others may be too durable, making it difficult to separate from the metal castings (increasing shakeout times). Based on the myriad of conditions impacting binder selection, there is no single binder system that will work in all applications. When considering a national standard, it may not be technically feasible to ban organic binder systems or otherwise require the use of only inorganic binder systems because there are some castings for which an inorganic binder system may not work due to the size, shape, or tolerance of the part to be cast or the climate (temperature and humidity) where the foundry is located. 
      During the development of the MACT standard, the EPA looked to ban selected high-HAP emitting binder formulations. In fact, the MACT standards include prohibition on the use of furan warmbox binder formulations that contain methanol as a specific ingredient of the catalyst portion of the binder system. When Subpart EEEEE was first proposed (67 FR 78274, December 23, 2002), it included:
 a prohibition on mold and core coatings that contain HAP as an ingredient of the liquid component of the coating formulation;
 a prohibition on the use of furan warmbox binder formulations that contain methanol as a specific ingredient of the binder system (both the resin component and the catalyst component);
 a requirement for foundries using phenolic urethane coldbox or nobake binders to use a chemical formulation in which the solvents were naphthalene-depleted; and
 a requirement for foundries using a binder system other than those outlined above to conduct a study to identify, evaluate and, if practical, adopt a reduced-HAP binder formulation.   
      Except for the catalyst component of the furan warmbox binder formulation, these prohibitions and requirements were not finalized due to information received during the public comment period. For example, emissions from a naphthalene-depleted phenolic urethane binder system were measured for PCS lines at the Castings Emission Reductions Program (CERP) facility, and HAP emissions actually increased compared to the traditional naphthalene -containing formulation. The use of the naphthalene-depleted solvent reduced HAP emissions during the mold- and core-making process, but when considering the entire casting operations, the naphthalene-depleted solvent formulations did not yield net reductions in HAP emissions. These findings indicate that a reduced-HAP binder formulation requirement is not necessarily equivalent to a low HAP-emitting binder system when considering both mold and core making and PCS emissions (i.e., the life cycle of the casting). Performance issues were also identified with the "no methanol" furan warmbox resin formulations in locations of high relative humidity. 
      Thus, the EPA had previously evaluated setting a national emissions standard requiring the use of low-HAP-emitting binder systems. However, the EPA concluded that it was not technically feasible or effective to develop nationwide bans on HAP-containing binder formulations. The EPA also found that low HAP-containing binder formulations did not necessarily translate into low HAP-emitting binder formulations and that it was difficult and expensive to assess the life-cycle emissions attributable to the use of these chemical binder systems. There have not been any significant changes in these chemical binder systems or in the ability to effectively assess the impact of binder system reformulations on the overall emissions from iron and steel foundries. As such, a national emission standard focused on low HAP-emitting binder systems may not be technically feasible. It may be possible, however, for some foundries to use binder system reformulations or switch to inorganic binder systems in order to comply with a potential organic HAP emissions limit for mold- and core-making operations and PCS lines.
      Use of inorganic (or other low HAP-emitting) binder systems may require some changes in the sand/binder mixing equipment used at the foundry, but the capital costs for converting binder systems is typically small. The cure times may be different, and longer cure times may require larger storage areas, which may be difficult and expensive to accommodate in a retrofit scenario. The low HAP-emitting binder systems are typically more expensive to use. The binder chemicals themselves may be more expensive, or their use may require a higher grade of sand, increasing operating costs. In cases where a low HAP-emitting binder system can be used to produce a given casting at a lower cost, foundry owners and operators would naturally select the low-cost alternative. Therefore, if a foundry has not yet switched to a low-emitting binder system, but subsequently switches its binder formulations to meet specific organic HAP emissions limits for mold and core making or PCS lines, we expect the foundry will have increased operational costs.    
3.2	Carbon Adsorption Systems
      Carbon adsorption is a conventional add-on control measure for controlling organic pollutant emissions. The EPA identified carbon adsorption for control of PCS lines during the development of the MACT standard for major source iron and steel foundries (67 FR 78292). Two foundries that used chemically bonded sand for molds and/or cores in automated or pallet PCS lines were identified as having carbon adsorption systems. Foundries that use high levels of chemically bonded sand tend to have higher organic emissions from their PCS lines. 
      The control efficiency for a carbon adsorption system is typically 90 to 95 percent. However, at low concentrations, the control efficiency generally declines, and the EPA has a long history of establishing an alternative organic concentration limit of 20 ppmv to address cases of low inlet concentrations. Based on the low organic HAP concentrations observed in measured emissions from well-captured PCS lines, the EPA established a volatile organic HAP limit of 20 ppmv for automated conveyor and pallet cooling lines and automated shakeout lines at new iron and steel foundries that use a sand mold system [40 CFR 63.7690(a)(10)] and did not provide a control efficiency alternative. Note that this control requirement was established for sources at new iron and steel foundries where close capture systems can be integrated into the foundry design. For existing iron and steel foundries, the EPA developed an operating requirement to develop procedures to provide an ignition source to sand mold vents at each pouring station or line in efforts to light off (combust) organic pyrolysis products released from the sand mold system [40 CFR 63.7720(b)(6)]. Thus, PCS lines at existing iron and steel foundries may have no capture system present to collect the gases emitted from the PCS lines. While not specifically required for mold- and core-making operations, carbon adsorption may also be applied to control organic emissions from these operations.
      There are two types of carbon adsorption systems: a single use carbon cannister system or a regenerable carbon adsorption system. A regenerable carbon adsorption system has multiple carbon beds. In a three-bed system, for example, two beds will be online adsorbing organics from the process gas stream being controlled and one bed will be off-line being regenerated, typically using steam. One advantage of a regenerable carbon adsorption system over a single use system (or a destructive control system such as a thermal oxidizer) is that the organic pollutants can be recovered potentially for beneficial use, which may offset some of the operational costs. For the control of PCS line emissions, the organic pollutants emitted are pyrolysis products, and it is unlikely that the organics recovered will have any value. For mold and core making, the organics emitted are more likely to be solvents used to carry the active ingredients of the binder formulation. During the mold- and core-making process, some portion of these inert solvents can evaporate. In this case, the organics recovered may be able to be recovered and recycled by the binder manufacturer. A regenerable carbon adsorption system will use less carbon over time than a carbon cannister system, but it will have steam costs (and secondary emissions from steam generation) and additional equipment costs.
3.3	Incineration/Thermal Oxidizers
      Incineration is also a conventional add-on control measure for controlling organic pollutant emissions. The EPA identified one foundry that operated a thermal oxidizer to control emission from their PCS line (67 FR 78292). This foundry is also expected to have had higher than average emissions due to the use of chemically bonded sand for molds and/or cores in automated or pallet PCS lines. 
      Incineration is typically 98 to 99.9 percent efficient at reducing organic emissions. However, like carbon adsorption, the control efficiency generally declines at low concentrations, and the EPA has a long history of establishing an alternative organic concentration limit of 20 ppmv for incinerators and thermal oxidizers. 
      Incineration can be used to control organic emissions from both mold- and core-making operations and PCS lines. However, for these sources, the organic content of the gas to be controlled is quite low, and the volume of gas to be treated is quite large. Additionally, the gas is emitted at or only slightly above room temperature. This means that natural gas or other fuel must be burned to heat the collected gas to incineration temperatures. The combustion of the auxiliary fuel generates additional emissions of carbon dioxide (CO2) and nitrogen oxides (NOx). Because of the amount of fuel needed to incinerate this dilute gas stream, incineration as a control measure for mold and core making and PCS lines may generate environmental disbenefits (in terms of CO2 and NOx emissions) that are similar in magnitude to the environmental benefits achieved by reducing organic HAP emissions.
4.	Control Cost Estimates for Organic HAP Emissions from Iron and Steel Foundries 
4.1	Low-emitting Binder System Costs
      It is difficult to assess the costs of converting to low-emitting binder systems, if a suitable low-emitting binder system can be identified. The costs would be dependent on the equipment and costs for the current binder system as well as the equipment and costs of the alternative binder system that may be suitable for the foundry's castings. As noted in our discussion of low-emitting binder systems in Section 3.1 of this memorandum, we generally expect that capital costs would be small because the equipment used to mix the binder with the sand can often be used for different binder formulations. Depending on the binder formulation, curing times, and other considerations, there could be significant capital costs associated with switching binder systems. We also expect that the alternative binder formulation would be more expensive to use on a per-castings-produced basis. However, without specific information on the different binder systems in use and that may be applicable, it is difficult to quantify these costs.
      In subsequent subsections, we provide cost estimates for add-on control systems. We do expect that there will be cases where a suitable alternative binder system could be used to meet the organic HAP emission limitation established for the add on control system. In cases where this substitution provides a lower compliance cost than an add-on control system, that foundry is likely to elect to use a low-emitting binder system to comply with the potential organic HAP emission limitation.  However, we cannot quantify the number and the costs associated with foundries that may be able to comply using low-emitting binder systems. 
4.2 	Model Emissions Stream
      Key parameters needed to develop the costs of carbon adsorption systems as well as thermal oxidizers include the volumetric flow rate and temperature of the gas stream to be treated, the concentration of specific organics present in the gas stream, and the number of hours the source is operated per year. 
      The volumetric flow rate of gas to be treated is dependent on the size of the source and the ability to provide a close capture hooding system. If the emission source can be enclosed or the capture system is located very near the source, less volume of gas is needed to ensure good capture of emissions. If a side draft hood or canopy hood well above the source is used, more air flow is needed for good capture. The amount of gas needed for good capture, however, does not impact the total load of organics to the control system. An enclosed system will have a lower exhaust flow rate with higher concentrations of organics, whereas the side draft hood system will have higher flow rates and lower concentrations (more dilution air entrained).  Because this analysis focuses on applying controls to existing sources, we expect that side draft hooding or canopy hooding somewhat removed from the source would be needed. Therefore, we estimated that all captured emissions would be quite diluted, on the order of 40 to 50 ppmv total organics similar to that observed in the CERP Mexico study (CERP, 1999a). 
      For this screening analysis, we developed a series of four model plants based on a range of organic HAP emissions reported for foundries in the 2014-based emissions inventory file used for the risk assessment (EPA, 2019). Based on emissions measured from PCS lines (CERP 1999a, 1999b, 2000), we estimated that approximately 65 percent of VOC emissions were also organic HAP, so VOC emissions estimated to be 1.54 times (1/0.65) the organic HAP emissions. We then developed four different-sized "model plants" to span the range of VOC emissions projected from the NEI data. The four model plants developed are summarized in Table 1. 
        Table 1. Model Plant Organic HAP Flow Rates and Operating Hours

                                Model Plant No.

                               Flow Rate (acfm)

                           Average VOC Conc.  (ppmv)

                            Annual Operating Hours

                        Maximum VOC Emissions (ton/yr)
                                       1
                                    90,000
                                      50
                                     4000
                                     120.5
                                       2
                                    60,000
                                      50
                                     3000
                                     60.3
                                       3
                                    30,000
                                      45
                                     3000
                                     27.1
                                       4
                                    20,000
                                      45
                                     2000
                                     12.1
      
      In this screening analysis, we used the aggregate emissions from mold and core making and PCS lines combined and developed a single control system for the aggregate emissions. In practice, these emission sources may be a large distance apart and it may not be practical to employ a single control system for the aggregate emissions. However, for a screening assessment, this assumption represents the most cost-effective control scenario. If the costs for the aggregate control system are not cost effective, we can conclude with certainty that separate control systems for mold and core making and PCS lines will not be cost effective.
      Each foundry was assigned to a model plant based on their projected VOC emissions from reported organic HAP emissions from mold- and core-making operations and PCS lines. If a facility had organic HAP emissions of 20 tons/yr, the VOC emissions were projected to be 30.8 tons/yr (20/0.65). Because these emissions are greater than the maximum VOC emission from Model Plant 2 and less than the maximum for Model Plant 3, this facility was assigned costs for Model Plant 3. If the projected VOC emissions from mold and core making and PCS lines were less than 1 ton/yr, we assumed the foundries would be able to meet the potential organic HAP emissions limit without the addition of an add on control device. Table 2 provides a listing of the foundries and emission estimates for foundries with projected VOC emissions greater than 1 ton/yr. 
Table 2. Summary of Organic HAP Emissions Estimates and Model Plant Assignments
                                 Facility Name
                                     City
                                     State
                   Cumulative Organic HAP Emissions (ton/yr)
                       Projected VOC Emissions (ton/yr)
                             Assigned Model Plant
Waupaca Foundry Inc
Tell City
IN
                                                                          74.18
                                                                         114.12
                                       1
JOHN DEERE FOUNDRY WATERLOO
Waterloo
IA
                                                                          58.10
                                                                          89.38
                                       1
Mueller Co
Albertville
AL
                                                                          55.56
                                                                          85.47
                                       1
WAUPACA FOUNDRY INC-PLANTS 2 / 3
Waupaca
WI
                                                                          51.63
                                                                          79.44
                                       1
Caterpillar Inc
Mapleton
IL
                                                                          29.73
                                                                          45.74
                                       2
WAUPACA FOUNDRY INC PLANT 1
Waupaca
WI
                                                                          27.67
                                                                          42.57
                                       2
WAUPACA FOUNDRY INC -PLANT 4
Marinette
WI
                                                                          24.28
                                                                          37.36
                                       2
AMERICAN CAST IRON PIPE COMPANY
Birmingham
AL
                                                                          19.91
                                                                          30.64
                                       2
ARDMORE FOUNDRY
Ardmore
OK
                                                                          18.43
                                                                          28.36
                                       2
Grede II LLC - Columbiana
Columbiana
AL
                                                                          15.98
                                                                          24.58
                                       2
NEENAH FOUNDRY CO - PLANTS 2 & 3 
Neenah
WI
                                                                          15.58
                                                                          23.97
                                       2
U. S. PIPE & FOUNDRY COMPANY, LLC
Bessemer
AL
                                                                          14.38
                                                                          22.12
                                       2
AMERICAN CASTING
Pryor
OK
                                                                          12.89
                                                                          19.83
                                       3
MAGOTTEAUX PULASKI
Pulaski
TN
                                                                          12.52
                                                                          19.26
                                       3
CLOW VALVE COMPANY - FOUNDRY
Oskaloosa
IA
                                                                          11.75
                                                                          18.08
                                       3
Tyler-Union Foundry Company
Anniston
AL
                                                                          10.79
                                                                          16.60
                                       3
AARROWCAST INC
Shawano
WI
                                                                          10.65
                                                                          16.39
                                       3
Liberty Casting Co 
Delaware
OH
                                                                          10.10
                                                                          15.53
                                       3
GREDE LLC - REEDSBURG
Reedsburg
WI
                                                                           8.41
                                                                          12.94
                                       3
QUALITY ELECTRIC STEEL CASTINGS
Houston
TX
                                                                           8.10
                                                                          12.47
                                       3
HOEGANAES CORPORATION
Gallatin
TN
                                                                           6.41
                                                                           9.86
                                       4
Nidec Minster Corporation 
Minster
OH
                                                                           4.56
                                                                           7.01
                                       4
Grede II LLC  Brewton
Brewton
AL
                                                                           3.66
                                                                           5.63
                                       4
Bradken - Atchison/St. Joseph
Atchison
KS
                                                                           3.57
                                                                           5.49
                                       4
Harrison Steel Castings Company
Attica
IN
                                                                           2.32
                                                                           3.57
                                       4
      
4.3 	Carbon Adsorption System Costs
      The capital investment and total annualized costs for carbon adsorption systems were developed for each of the model plant emissions using the recently updated chapter of the EPA Air Pollution Control Cost Manual (U.S. EPA, 2018). Costs for a three-bed regenerable system were developed. Costs for non-regenerable carbon cannisters were also estimated, but the costs were significantly higher than the regenerable system, so only the costs of the regenerable systems are provided in this memorandum. The adsorption characteristics were based primarily on benzene: a "k" parameter of 0.6 was used and an "m" parameter of 0.23 was used. Capital costs were escalated to 2017 dollars, and a retrofit factor of 1.3 was applied to the capital costs to account for additional costs associated with retrofitting the control system to an existing plant. Capital costs were annualized using a 5 percent annual interest rate. The default equipment life of 15 years was used; we assumed carbon life of 2 years. Operating costs were based on 2017 values summarized in Table 3.
               Table 3. Summary of 2017 Utility and Labor Rates
      Description
   Cost
    Units
 Electricity 
 $0.0688
 per kWh
 Steam 
 $7.70
 per 1,000 lbs
 Natural Gas 
 $4.00
 per 1,000 scf
 Cooling Water 
 $3.55
 per 1,000 gallons of water
 Operator Labor Rate
 $27.48
 per hour
 Maintenance Labor Rate
 $30.23
 per hour
 Carbon Cost 
 $3.80
 per lb
 Re-Sale Value of Recovered VOC 
 $0.00
 per lb
 Disposal/Treatment Cost for Recovered VOC 
 $0.11
 per lb
      
      Auxiliary equipment costs for hooding, dampers, fans and stacks were estimated to be $50,000 for all model plants. The model plant control costs are summarized in Table 4. These costs were then used to estimate the nationwide control costs based on the number of foundries assigned to each model plant as provided in Table 5.
       Table 4. Cost Estimates for Model Plant Carbon Adsorption Systems
                                Model Plant No.
                            Total Capital Cost ($)
                         Total Annualized Cost ($/yr)
                            Steam Use (1,000 lb/yr)
                          Electricity Use (kW-hr/yr)

                                       1

                                   1,589,000

                                    380,300

                                      874

                                    232,300

                                       2

                                   1,261,000

                                    274,700

                                      422

                                    117,200

                                       3

                                    860,800

                                    179,600

                                      190

                                    56,040

                                       4

                                    707,700

                                    135,100

                                      84

                                    25,440
      
Table 5. Development of Nationwide Cost Estimates for Carbon Adsorption Systems
                                Model Plant No.
                           No. of Foundries Assigned
                       Aggregate Total Capital Cost ($)
                    Aggregate Total Annualized Cost ($/yr)
                                       1
                                       4
                                   6,356,000
                                   1,521,000
                                       2
                                       8
                                  10,088,000
                                   2,198,000
                                       3
                                       8
                                   6,886,000
                                   1,436,000
                                       4
                                       5
                                   3,539,000
                                    676,000
                                  Nationwide
                                      25
                                  26,870,000
                                   5,830,000
      
      The carbon adsorption systems were designed to achieve 90 percent emission reduction. The emission reduction was determined based on the actual emissions projected for each facility. The aggregate organic HAP and VOC emissions from Table 2 are 511 and 786 tons/yr, respectively. Assuming 90 percent emission reductions, the nationwide organic HAP and VOC emissions reductions for sources in the 40 CFR part 63, subpart EEEEE, source category are estimated to be 460 and 708 tons/yr, respectively. 
4.4 	Thermal Oxidizer Costs
      The capital investment and total annualized costs for thermal oxidizer control systems were developed for each of the model plants using the recently updated chapter of the EPA Air Pollution Control Cost Manual (U.S. EPA, 2017). Costs were developed for 70% recuperative thermal oxidizers as well as regenerative thermal oxidizers. The capital costs for recuperative thermal oxidizers are much lower than regenerative thermal oxidizers, but the operating and annualized costs are higher. Costs for catalytic thermal oxidizers fell between the costs of the recuperative and regenerative thermal oxidizers, so we are only reporting the recuperative and regenerative system costs to provide a range for all thermal oxidizer systems. We used 90 ºF as the inlet gas temperature and operating temperatures of 1,600 ºF and 1,900 ºF for the recuperative and regenerative thermal oxidizers, respectively. We assumed a pressure drop of 20 inches of water for both systems. Capital costs were escalated to 2017 dollars, and a retrofit factor of 1.3 was applied to the capital costs to account for additional costs associated with retrofitting the control system to an existing plant. Capital costs were annualized using a 5 percent annual interest rate. The default equipment life of 20 years was used. Operating costs were based on 2017 values summarized in Table 3.
      The model plant control costs for recuperative and regenerative thermal oxidizers are summarized in Tables 6 and 7, respectively. These model plant costs were then used to estimate the nationwide control costs based on the number of foundries assigned to each model plant (provided previously in Table 5). These nationwide costs are also provided in Tables 6 and 7. 
 Table 6. Cost Estimates for Model Plant Recuperative Thermal Oxidizer Systems
                                Model Plant No.
                            Total Capital Cost ($)
                         Total Annualized Cost ($/yr)
                        Natural Gas Use (1,000 scf/yr)
                          Electricity Use (kW-hr/yr)
                                       1
                                   1,436,000
                                   1,424,000
                                    279,400
                                   1,404,000
                                       2
                                   1,298,000
                                    274,700
                                    139,700
                                    702,000
                                       3
                                   1,091,000
                                    179,600
                                    69,850
                                    351,000
                                       4
                                    98,300
                                    135,100
                                    31,110
                                    156,000
                                  Nationwide
                                  29,800,000
                                  17,100,000
                                   2,950,000
                                  14,820,000
      
 Table 7. Cost Estimates for Model Plant Regenerative Thermal Oxidizer Systems
                                Model Plant No.
                            Total Capital Cost ($)
                         Total Annualized Cost ($/yr)
                        Natural Gas Use (1,000 scf/yr)
                          Electricity Use (kW-hr/yr)
                                       1
                                   4,265,000
                                    815,500
                                    42,290
                                   1,404,000
                                       2
                                   3,091,000
                                    532,200
                                    21,150
                                    702,000
                                       3
                                   2,659,000
                                    406,900
                                     8,837
                                    351,000
                                       4
                                   1,157,000
                                    231,500
                                     4,702
                                    156,000
                                  Nationwide
                                  70,700,000
                                  11,900,000
                                    432,500
                                  14,820,000
5.	Summary
      While thermal oxidizers generally achieve higher control efficiencies than carbon adsorption systems, given the relatively low starting concentrations, we expect both systems to comply with the lower concentration limit of 20 ppmv. Consequently, we do not expect an appreciable difference in the control efficiencies between thermal oxidizers and carbon adsorption in this application. Table 8 summarizes the cost effectiveness of the organic HAP control systems for major source iron and steel foundries for the add-on control systems evaluated. Based on a comparison of the total annualized costs, we expect most foundries would elect to use a carbon adsorption system rather than thermal oxidizers. Some foundries may be able to comply with an organic HAP emission limit for mold and core making and PCS lines using low-emitting binder formulations, which would likely reduce compliance costs. However, we also expect some foundries may be designed such that one control system may be needed for mold and core making and a separate control system may be needed for PCS line emissions, which would increase compliance costs.   
      
   Table 8. Comparison of Cost-Effectiveness for Organic HAP Control Systems
                                Control System
                       Nationwide Total Capital Cost ($)
                    Nationwide Total Annualized Cost ($/yr)
                   Organic HAP Emission Reduction (tons/yr)
                Cost Effectiveness ($/ton organic HAP reduced)
                               Carbon Adsorption
                                  26,870,000
                                   5,830,000
                                      460
                                    12,700
                         Recuperative Thermal Oxidizer
                                  29,800,000
                                  17,100,000
                                      460
                                    37,100
                         Regenerative Thermal Oxidizer
                                  70,700,000
                                  11,900,000
                                      460
                                    25,900
      
6.	References
CERP. 1999a. Foundry process emission factors: baseline emissions from automotive foundries in Mexico. January 19. 
CERP. 1999b. Baseline testing emission results pre-production foundry. November 1. 
CERP. 2000. Baseline testing emission results production foundry. February 7. 
U.S. Environmental Protection Agency. 2017. EPA Air Pollution Control Cost Manual. 7th Edition. Incinerators and Oxidizers. Section 3.2, Chapter 2. November. Available at:  https://www.epa.gov/sites/production/files/2017-12/documents/oxidizersincinerators_chapter2_7theditionfinal.pdf
U.S. Environmental Protection Agency. 2018. EPA Air Pollution Control Cost Manual. 7th Edition. Carbon Adsorbers. Section 3.1, Chapter 1. October. Available at: https://www.epa.gov/sites/production/files/2018-10/documents/final_carbonadsorberschapter_7thedition.pdf
U.S. Environmental Protection Agency. 2019. Residual Risk Assessment for Iron and Steel Foundries Source Category in Support of the 2019 Risk and Technology Review Proposed Rule, Appendix 1. In Docket ID No. EPA-HQ-OAR-2019-0373.
