U.S. EPA Proposes Benzene Fenceline Monitoring for Refineries Amongst Other MACT and NSPS Changes

U.S. EPA has proposed an amendment to 40 CFR Parts 63, Subpart CC and Subpart UUU that requires all petroleum refineries to deploy passive time-integrated benzene samplers at the fenceline (where fenceline is equivalent to the facility property line) of their facility to alert them of fugitive emissions from leaks of Subpart CC and UUU process equipment.  The proposed amendment also requires the operation of an on-site meteorological monitoring system in conjunction with the fenceline benzene monitoring.  U.S. EPA proposed the use of 12 to 24 passive samplers placed at 15 to 30 degree intervals along the perimeter of each refinery, depending on the size and shape of the property.  Ideally, for a small refinery (< 750 acres), samplers would be placed at 30 degree intervals for a total of 12 samplers.  For a medium-sized refinery (750 to 1,500 acres) a total of 18 samplers would be deployed at 20 degree intervals, and for a large refinery (> 1,500 acres), samplers would be distributed at 15 degree intervals for a total of 24 samplers.  An extra sampling location may be required when there is an emission sources located close to the fenceline/property line.  A sampling event would consist of 14 days.

Passive Time-Integrated Diffusive Tube Samplers

Passive time-integrated diffusive tube samplers involve a series of tube samplers at set intervals along the fenceline that measure the ambient air concentration at each sampling location.  The samplers consist of a small tube filled with an adsorbent and capped with a special cover with small holes to allow the ambient air to diffuse into the tube at a small, fixed rate.  Passive diffusive tube samplers are currently used by the European Union for benzene monitoring and were determined by U.S. EPA to be a cost-effective technique for measuring benzene concentrations based on a year-long ambient monitoring trial at a Texas refinery.  The fenceline monitoring pilot project took place at Flint Hills West Refinery in Corpus Christi, TX in 2009.  The project demonstrated that the low-cost, high spatial density monitors could be used by modestly trained personnel, and provided useful, timely facility diagnostics, such as alerts of fugitive leaks, when combined with on-site meteorological measurements.  The first year cost for installation and equipment could reach up to $100,000 for a large refinery with 24 sampling locations, with an annualized cost of about $40,000 per year for maintenance and analyses.  However, these costs do not include effort associated with development of a monitoring plan, routine recordkeeping, and semi-annual reporting.

Benzene Concentration Action Level

U.S. EPA modeled fenceline benzene emissions from the 2011 refinery information collection request (ICR) to determine an appropriate benzene concentration action level, the concentration that, if exceeded, a refinery would need to take corrective action.  Based on U.S EPA’s air dispersion modeling study, the maximum post-MACT control benzene concentration modeled at the fenceline was determined to be 9 micrograms per cubic meter (µg/m3) of benzene.  This concentration reflects refinery fugitive emission sources only and excludes a background concentration.  A facility would exceed the concentration action level when the highest of the rolling annual average fenceline concentrations corrected for background is greater than 9 µg/m3.

Refineries would have to calculate the rolling annual average within 30 days of completion of each sampling episode.  If the rolling annual average exceeds the concentration action level, within five (5) days of calculating the rolling annual average, the refinery must initiate a root cause analysis to determine the primary and contributing causes of the exceedance.  This root cause analysis must be completed within 45 days of initiation, followed by the corrective action plan to reduce emissions.  The refinery would not be deemed out of compliance with the proposed concentration action level, as long as appropriate corrective action measures are taken.

The recordkeeping and reporting of this data would consist of electronically reporting the fenceline data at each monitor within 45 days of the end of semiannual periods.  The U.S. EPA is proposing to submit the raw data as well as the highest fenceline concentration data through its public electronic reporting and data retrieval portal.  This submittal will include the individual sample results from each monitor, the coordinates of all monitors in decimal degrees, the biweekly 12-month rolling average concentration difference values, and notes for each value (e.g. background corrections used, if the value was under detection, or an outlier that was removed from the data set).

Site-Specific Ambient Monitoring Plan

Sources that have local sources of emissions will likely need to develop a site-specific ambient monitoring plan to address how background concentrations will be determined and to provide a description of any unique features to a source’s monitoring program (e.g., the gradual scale back of sampling for situations with extremely low monitored fugitive emissions are measured).  Since only isolated sources are likely to avoid the prospect of background emissions affecting the fenceline/property line measurements, it will be critical that sources prepare a site-specific monitoring plan.  The monitoring plan must be submitted to U.S. EPA for review and approval.

So What Does This Mean for Petroleum Refineries in the U.S.?

Existing refineries would be required to deploy samplers no later than three (3) years after the effective date of the final rule and new sources would be required by startup or the effective date of the rule, whichever is later).  Refineries would be required to demonstrate compliance with the concentration action level one (1) year following the compliance date since the concentration action level is determined by a one (1) year rolling average.

The U.S. EPA is requesting comments on the following topics regarding the proposed benzene monitoring requirement:

• Suitability of selecting benzene or other HAP, including PAH or naphthalene, as the indicator to be monitored.
• Whether it is appropriate to require multiple HAP to be monitored at the fenceline/property line considering capital and annual cost for additional monitors, and if multiple pollutants should be monitored, which pollutants.
• The location/adaptation of the monitoring requirements for facilities that have fugitive emission sources located on property that is not contiguous.
• Alternative approaches for determining background concentrations.
• How to identify conditions where the lowest mean concentration (LMC) differential may underestimate the highest fenceline concentration and when to use the LMC versus the uniform background concentration.
• How to adjust for background contributions for complex cases.
• Adequacy of the 9 µg/m3 benzene concentration action level or approach to determine the concentration action level.
• Whether it is appropriate to establish a standard time frame for compliance with actions listed in a corrective action plan, and whether the approval of the corrective action plan should be delegated to state, local, and tribal governments.
• Whether the corrective action requirements should be limited to exceedances of the fenceline concentration solely from refinery emission sources and whether a refinery owner should be allowed to exceed the concentration action level if it is possible to demonstrate the exceedance is due to a nonrefinery emission source.
• Suggestions for other ways U.S. EPA can monitor and improve this requirement.
• Approaches for reducing or eliminating the fenceline monitoring requirements for facilities that consistently measure fenceline concentrations below the concentration action level – such as minimum time period required, level of performance, etc.

As refineries begin to plan for the benzene fenceline monitoring, don’t hesitate to ask for assistance.  ALL4 would be glad to discuss any of the following:

• Installation of passive samplers, or other monitoring equipment.
• Training of local personnel on equipment.
• Creation of spreadsheet-like tools to record bi-weekly data and calculate annual rolling averages.
• Analyzing the raw data to subtract out background concentrations as appropriate.
• Installation and maintaining new/upgraded meteorological equipment for on-site ambient meteorological data (will help with background concentration calculations).
• Conducting a quality assurance audit of on-site meteorological monitoring system.
• Preparing the semi-annual electronic reports for U.S. EPA.
• Preparing root-cause analysis and corrective action plans to submit to U.S. EPA if refinery exceeds concentration action level.
• Developing a monitoring plan to develop an alternate method for quantifying background concentrations.

Additional Amendments to the Refinery MACT 1 and 2

Besides the flare changes discussed in the ALL4’s May 2014 4 The Record and the benzene fenceline monitoring requirements discussed above, there are four (4) other noteworthy proposed actions that are included in the revisions to the Refinery MACT 1 (Subpart CC) and MACT 2 standards (Subpart UUU).  Most important among the four (4) proposed actions is the startup, shutdown and malfunction (SSM) revisions to the MACT 1 and 2 rules.  The remaining three (3) revisions include the following:

• Updating the General Provisions Tables in MACT 1 and 2 to reflect the SSM vacatur and to update cross-reference citations.
• Adding electronic reporting requirements in MACT 1 and 2.
• Clarifying requirements related to open-ended valves or lines.

For each of the items U.S. EPA is requesting comments from interested stakeholders.

In addition to the changes proposed to Subparts CC and UUU, U.S. EPA has also added amendments to the Refinery New Source Performance Standards (NSPS) J and Ja.  A brief summary of these proposed changes is provided since Subpart UUU references Subpart J and both Subparts J and Ja apply to petroleum refineries.

SSM

In 2008, the Court of Appeals for the District of Columbia Circuit ruled that emission standards under Section 112 of the Clean Air Act (CAA) apply at all times.  Based on this ruling, the SSM provisions in Subparts CC and UUU no longer apply.  Therefore, U.S. EPA is proposing alternate emission standards for start-up and shutdown conditions for a few select sources as noted below.  U.S. EPA believes that with the exception of the sources noted below, all other emission units, including wastewater treatment, should be able to comply with existing emission limits during start-up and shutdown.

• Fluid catalytic cracking unit (FCCU) controlled by an electrostatic precipitator (ESP) during start-up with torch oil.
• FCCU during start-up.
• Sulfur recovery unit (SRU) during shutdown.

When a FCCU controlled with an ESP begins start-up on torch oil, the ESP is offline.  U.S. EPA believes that cyclone separators internal to the FCCU are sufficient to control PM and metal HAPs during start-up.  As a result, U.S. EPA is proposing a six (6)-minute rolling average 30% opacity limit during the period that torch oil is utilized.  Certification with the opacity limits may be performed with visual opacity readings.

Emissions of organic HAPs from an FCCU are either controlled by a post-combustion pollution control device or by the FCCU itself.  U.S. EPA anticipates that for an FCCU that is equipped with a post-combustion device, FCCU organic HAP emissions during start-up will be adequately controlled by simply operating the air pollution control device.  For an FCCU without an air pollution control device, U.S. EPA recognizes that the FCCU acts to control its own HAP emissions, even during periods of start-up.  Thus, U.S. EPA proposes that an FCCU without a post combustion control device, meet an excess oxygen (O2) concentration operating limit of at least 1% by volume on a dry basis and averaged on a 1-hour basis during start-up.

Finally, changes to the shut-down provisions for select SRU emissions units are proposed.  Specifically for SRU units that elect to/are required to comply with the 250 ppm sulfur dioxide (SO2) emission limit associated with NSPS Subpart J, the proposed shut-down provisions establish a 300 ppm total reduced sulfur (TRS) emissions limit.  Compliance with the 300 ppm TRS emission limit is achieved by operating the thermal oxidizer at 1,200 °F and a minimum O2 concentration of 2% by volume on a dry basis or by diverting the SRU gases of a flare meeting specific design criteria.

U.S. EPA states in their proposed rule revisions that all other emissions units regulated under Subpart CC and UUU do not need to have revised emission limits to reflect start-up and shut-down conditions.  U.S. EPA expects that exiting air pollution control equipment will function the same during normal operations as during start-up and shut-down.

Finally, U.S. EPA proposed that the inclusion of a regulatory affirmative defense provision is not appropriate given the April 2014 decision by the Court of Appeals for the District of Columbia.  The Court decision eliminated the protection from civil penalties for malfunction events where an emission limit is exceeded.  However, facilities can still present arguments to mitigate the amount of civil penalties, which are levied by the court system, and also to U.S. EPA administrative penalties, which are levied by U.S. EPA.

Changes to Applicability of NESHAP General Provisions

The removal of the SSM provisions in Subparts CC and UUU also required revising the summary of the applicable NESHAP General Provisions that apply to the respective Subparts.  The NESHAP General Provisions include multiple requirements for which affected sources must demonstrate compliance.  Since these General Provisions related to SSM would no longer apply, Table 6 for Subpart CC and Table 44 for Subpart UUU are proposed for revisions.  The revisions to Tables 6 and 44 address the following:

• General duty provisions to minimize emissions during SSM are no longer applicable.
• Compliance with opacity and non-opacity standards during SSM events is not exempted.
• Eliminate the preparation/implementation of an SSM Plan.
• Performance Testing cannot occur during periods qualifying as SSM and must occur during normal operation for which documentation is provided.
• Monitoring is required to show that good air pollution control measures are being practiced in lieu of maintaining an SSM Plan
• Recordkeeping within Subparts CC and UUU is established for periods of SSM to quantify emissions, differentiate between start-up and shut-down, show actions taken to minimize emissions, document the date, time, duration, cause, sources affected, estimates of each regulated pollutant emitted over an emission limits and how the emissions were calculated.  Monitoring associated with the SSM plans is no longer required.
• Reporting formerly required as part of the SSM Plan is replaced with periodic reporting to submit the information developed as part of the new recordkeeping requirements.

In addition to the implementing the vacatur of the SSM provisions, Table 6 (Subpart CC) and Table 44 (Subpart UUU) are being updated to reflect other changes to the General Provisions.  These additional changes to the two tables include correcting certain cross-references and incorporating additional sections of the General Provisions that implement other subparts that are referenced by Subparts CC and UUU.

Electronic Reporting

Performance tests that incorporate test methods that are supported by the Electronic Reporting Tool (ERT) must submit the results electronically to U.S. EPA.  The results must be submitted with the Compliance and Electronic Reporting Interface (CEDRI) system through the Central Data Exchange (CDX).  Fenceline ambient air monitoring data would also be reported via CEDRI on a semi-annual basis.

Technical Amendments

U.S. EPA is proposing minor technical amendments to Subparts CC and UUU.  Under Subpart CC, U.S. EPA is providing clarification for assessing how open-ended valves and lines are considered to be sealed for compliance purposes.  U.S. EPA is proposing to establish a no detectable emissions level that is equivalent to an instrument reading of 500 ppm or less as measured at the cap, blind flange plug, or second valve to the open-ended valve.

Amendments to the Refinery NSPS J and Ja

NSPS Subparts J and Ja regulate criteria pollutants from petroleum refineries.  As part of the proposed NESHAP rulemaking, U.S. EPA is proposing changes to Subparts J and Ja.  The proposed changes are the result of several years of discussion with stakeholders.  The primary intent associated with the NSPS changes is to make the NESHAP and NSPS requirements consistent with respect to delayed coking units (DCU) and to correct and clarify other J and Ja requirements.

With respect to a DCU, the depressurization work practice requirements of NESHAP Subpart CC are more stringent than the depressurization work practice requirements of Ja.  U.S. EPA is requesting comment on administrative approaches for aligning the requirements of Subpart Ja with those of Subpart CC.

U.S. EPA is also proposing clarifying changes to the Subpart Ja applicable SO2 and TRS emission limits for sulfur recovery plants where oxygen enrichment is used.  Also, emissions averaging may be applied to the various emissions units within the specific sulfur recovery plant.

In addition to the changes mentioned above, U.S. EPA is proposing 15 changes to Subparts J and Ja.  Three (3) important changes are noted below.

• Amendments to 40 CFR Part §60.106a regarding monitoring with respect to O2 enrichment, emissions averaging, and adding “reduced sulfur compounds” when referring to select emissions limits
• Amendments to 40 CFR Part §60.102a that clarify the carbon monoxide (CO) boilers that are part of the FCCU can also be fuel gas combustion devices.  If a source meets the definition of more than one affected facility there may be multiple requirements that apply to the source.
• Amendments to 40 CFR Part §60.104a are being proposed to clarify that performance testing be performed once a calendar year with at least 8 months and no more than 16 months between testing.

The “Benzene Fenceline Monitoring for Refineries amongst other MACT and NSPS Changes” is just one element of U.S. EPA’s recent proposal.  Refer to Kristin Gordon’s blog for an overview of the proposal.  ALL4’s Roy Rakiewicz authored “U.S EPA Proposes Substantial Refinery Flare Operating and Monitoring Requirements” in May’s 4 The Record.  For more information on Benzene Fenceline Monitoring please contact Dan Dix at ddix@all4inc.com or 610.933.5246 x118.

Risk Assessment: Refinery MACT 1 and 2

Background

Section 112(f)(2) of the Clean Air Act (CAA) requires the U.S. EPA to determine for source categories subject to Maximum Achievable Control Technology (MACT) standards whether the emission standards provide an ample margin of safety to protect public health.  This assessment is carried out by the U.S. EPA in a two-step process that includes the primary objective of developing regulations to address any residual risk after implementation of the original Refinery MACT 1 and MACT 2 emission standards and work practices.  Following the risk assessment, the agency performs the prescribed second step of determining whether further requirements are necessary to provide an “ample margin of safety” under its evaluation of the original MACT rules.

The first step in the process of evaluating residual risk is to make a determination of what is an acceptable risk.  If the risks are not acceptable, then the U.S. EPA cannot consider cost in identifying the emissions standards necessary to bring the risks down to an acceptable level.  The second step is the determination of whether regulatory standards must be further strengthened in order to provide the required “ample margin of safety” necessary to protect public health.  The ample margin of safety is the level at which the standards must be set, unless an even more stringent standard is necessary to prevent an adverse environmental impact, taking into consideration costs, energy, safety and other relevant factors.

Risk Assessment

The determination of acceptable risk begins with a presumptive limit on the maximum individual lifetime risk (MIR) of approximately 1-in-10 thousand.  The MIR risk assessment considers the risk from both cancer and non-cancer exposures and considers acute exposures as well as chronic effects from life-time exposures.  Both long-term and short-term inhalation exposure concentrations and health risks from the petroleum refinery sector were addressed by the U.S. EPA in their MACT review by using the Human Exposure Model (Community and Sector HEM-3 version 1.1.0). 

The U.S. EPA evaluated risk using screening estimates of acute exposures and risks for each of the hazardous air pollutants (HAPs) of concern at the point of highest off-site exposure for each facility by assuming that a person is located at this spot at a time when both the peak (hourly) emissions rate and worst case dispersion conditions would occur.  The U.S EPA also conducted the screening analysis examining the potential for significant human health risks due to exposures via routes other than inhalation, such as by consumption of locally grown food.  They first determined whether any sources in the source category emitted any hazardous air pollutants known to be persistent and bio-accumulative in the environment (Lead(PB)-HAP).  For the petroleum refinery source categories, the U.S. EPA identified emissions of cadmium compounds, chlorinated dibenzodioxins and furans, lead compounds, mercury compounds, polychlorinated biphenyls, and polycylic organic matter.

In addition to assessing baseline inhalation risks and screening for potential multipathway risks, the U.S. EPA estimated risks considering the potential emission reductions that would be achieved by the different control options that they were considering for the MACT reassessment.  The U.S. EPA used the same emissions inventory that it used for their risk screening modeling and then applied emission reduction estimates for the different control options under consideration in order to calculate the post-control risk values.

The results of the chronic baseline inhalation cancer risk assessment indicated that, based on estimates of current actual emissions, the maximum individual lifetime cancer risk (MIR) posed by the refinery source category is 60-in-1 million, with benzene and naphthalene emissions from equipment leaks and storage tanks accounting for 98 percent of the MIR risk.  The total estimated cancer incidence from refinery emission sources, based on actual emission, levels is 0.3 excess cancer cases per year or one case in every 3.3 years.  When considering the MACT-allowable emissions, the maximum individual lifetime cancer risk was estimated to be up to 100-in-1 million, driven by emissions of benzene and naphthalene from refinery fugitives (e.g., storage tanks, equipment leaks and wastewater) and the estimated cancer incidence is estimated to be 0.6 excess cancer cases per year or one excess case in every 1.5 years.

The maximum modeled chronic non-cancer target organ-specific hazard index (TOSHI) value for the source sector, based on actual emissions, was estimated to be less than 1.  When considering MACT-allowable emissions, the maximum chronic non-cancer TOSHI value was estimated to be about 1.  All other HAP from the facilities modeled had worst-case acute hazard quotient values of less than 1, indicating that the HAP emissions are believed to be without appreciable risk of acute health effects.

To better understand the uncertainty of its multipathway screening analysis to either under predict or over predict risk, the U.S. EPA conducted a refined multipathway case study using the Total Risk Integrated Methodology (TRIM) TRIM.FaTE risk model for one of the petroleum refineries for which actual emissions data had been obtained.  The refinery was selected because of its nearby lakes and farms which provided a real lifetime exposure scenario, as well as the fact that the refinery was found to have one of the highest potential multipathway risks for PAH based on its Tier II risk analysis.  The refined analysis for this facility showed that the Tier II screening that had been performed for each pollutant at the site actually had over predicted the potential risk when compared to the refined analysis results.

Overall, the refined risk analysis predicted a potential lifetime cancer risk of 4-in-1 million for the MIR.  The non-cancer hazard quotient (HQ) was predicted to be well below 1 for all of the target organs.  The chronic inhalation cancer risk assessment estimated inhalation cancer risk around this same facility was found to be approximately 10-in-1 million, due in large part to emissions of naphthalene and 2-methylnaphthalene.  The U.S. EPA decided to look at these two results as additive, although they acknowledged it was highly unlikely, such that the person in the risk assessment with the highest chronic inhalation cancer risk was also assumed to be the same person with the highest individual multipathway cancer risk.  Therefore, the U.S. EPA combined these results to form the worst-case MIR for that facility which theoretically could be a 10-in-1 million risk.  While this refined assessment was performed on only a single facility, the results of this single refined analysis indicate that if refined analyses were performed for other sites, the risk estimates would consistently be lower than those estimated by their Tier II analysis.  In addition, the risks predicted by the multipathway analyses at most facilities are considerably lower than the risk estimates predicted by the inhalation assessment, indicating that the inhalation risk results are in all likelihood the primary factor that is causing any residual risk for petroleum refineries.

Summary – Environmental Risk Assessment Screening

In the Tier II screening analysis for PB-HAP, none of the individual modeled concentrations for any facility in the source category exceeded any of the ecological benchmarks.  For lead compounds, the U.S. EPA did not model any exceedances of the secondary lead National Ambient Air Quality Standards (NAAQS).  Therefore, the U.S. EPA did not conduct further risk assessment for lead compounds.

For acid gases, the average modeled concentration around each facility did not exceed any ecological benchmark.  In addition, for both hydrochloric acid (HCL) and hydrogen fluoride (HF), each individual concentration was below the ecological benchmarks for all facilities.  Therefore, the U.S. EPA did not conduct further risk assessment for acid gases.

The risk assessment results indicate that both the actual and allowable inhalation cancer risks to the individual most exposed are no greater than approximately 100–in-1 million, which is the presumptive limit of acceptability.  The MIR based on actual emissions is 60-in-1 million, approximately 60 percent of the presumptive limit.  Based on the results of the refined site specific multipathway analysis the U.S. EPA conclude that the ingestion cancer risk to the individual most exposed is significantly less than 100-in-1 million (1-in-10 thousand). In addition, the maximum chronic non-cancer TOSHI due to inhalation exposures is less than 1, and the U.S. EPA’s refined multipathway analysis indicates that non-cancer ingestion risks are estimated to be less than non-cancer risk from inhalation.  Finally, while the evaluation of acute non-cancer risks was very conservative, it also showed acute risks below a level of concern.  The U.S. EPA estimated that their proposed new standards would not affect the MIR, and do not by themselves justify additional emission reductions, but would if implemented would reduce the source category estimated cancer incidence by 15 percent. 

Conclusion

In conclusion, the U.S. EPA determined that the risks remaining after implementation of the existing Refinery MACT 1 and 2 rules were acceptable and did already provide “ample margin of safety” necessary to protect public health.

For more information about the proposed Refinery MACT revisions, refer to ALL4’s May 2014 and June 2014 4 The Record articles.

Permit Aggregation Ruling

For some time, U.S. EPA has considered the concepts of contiguous or adjacent, common control, and industrial groupings when determining whether emissions from multiple “sites” should be aggregated for air permitting purposes.  In a 2009 memo (“McCarthy Memo”) from U.S. EPA Administration agency air chief Gina McCarthy, the concept of “functional interrelatedness” was introduced as a significant consideration for permit writers when determining whether emissions should be aggregated.  U.S. EPA applied the functional relatedness test when determining that gas sweetening plants and related wells operated by Summit Petroleum were a single source for air permitting purposes. In Summit Petroleum Corp. v EPA et al., the 6^th Circuit ruled that physical proximity is the factor that determines if sites are “adjacent,” thereby negating U.S. EPA’s consideration of functional interrelatedness as a factor for single source determinations. 

In a December 2012 memo from U.S. EPA’s Office of Air Quality Planning and Standards, U.S. EPA opined that the 6^th Circuit decision in the Summit case applied only to Title V or New Source Review aggregation decisions under the jurisdiction of the 6^th Circuit (Michigan, Ohio, Tennessee, and Kentucky).  This memo established that U.S. EPA could and would use different criteria in different regions, specifically as it relates to the issue of “adjacent”.  Of course, given the complexities of the issue, the idea of differing criteria depending on geographical location did not sit well with the regulated community. 

On May 30, 2014 the U.S. Court of Appeals for the District of Columbia Circuit vacated U.S. EPA’s December 2012 memo that limited the non-inclusion of the functional interrelatedness consideration to only the area under the jurisdiction of the 6^th Circuit.  Absent further court action, this ruling means that U.S. EPA will no longer be able to use functional interrelatedness as a consideration for whether sites are considered adjacent for air permitting purposes.