General Overview of Past and Recent EPA Flare Enforcement Actions (and flare assessment techniques)

The recent 2012 Flare Consent Decrees (CD) issued by the U.S. Environmental Protection Agency (EPA) have been precipitated by alleged flare over-steaming practices in most all industrial applications. The impact of over-steaming on the subsequent degradation of combustion efficiency was revealed in 2005 when the ambient air monitor at the Meredith Hitchens Elementary School in Addyston Ohio began reporting significantly higher than background levels of 1,3 butadiene. The ensuing investigation eventually concluded that one of the flares at the nearby INEOS Lanxess Butadiene Plant was being significantly over-steamed, resulting in atmospheric emissions of butadiene at rates which were magnitudes of order greater than previously estimated. An investigation led by Texas Commission on Environmental Quality (TCEQ) - conducted between 2003-2006 in the Houston Brazoria counties - also focused on flares and it also highlighted the need to seriously address industrial flare operations. Over-steaming became and remains an area of intense interest. A further technical study commissioned by the TCEQ in 2008 further provided quantitative analysis on the significance of over-steaming practices and clearly demonstrated that air or steam-assisted flares can easily be "over assisted" resulting in very low combustion efficiencies. The TCEQ study clearly demonstrated the complex and inter-related impacts of multiple variables on flare performance, indicating that the management of flare combustion efficiency is a highly dynamic process. The EPA's response to this growing body of information has been enforcement first, with no regulatory improvement to date.

P a g e | 1 General Overview of Past and Recent EPA Flare Enforcement Actions General Overview of Past and Recent EPA Flare Enforcement Actions (and flare assessment techniques) Presented by: Troy M. Boley, Ph.D Introduction to Over-Assistance The recent 2012 Flare Consent Decrees (CD) issued by the U.S. Environmental Protection Agency (EPA) have been precipitated by alleged flare over-steaming practices in most all industrial applications. The impact of over-steaming on the subsequent degradation of combustion efficiency was revealed in 2005 when the ambient air monitor at the Meredith Hitchens Elementary School in Addyston Ohio began reporting significantly higher than background levels of 1,3 butadiene. The ensuing investigation eventually concluded that one of the flares at the nearby INEOS Lanxess Butadiene Plant was being significantly over-steamed, resulting in atmospheric emissions of butadiene at rates which were magnitudes of order greater than previously estimated. An investigation led by Texas Commission on Environmental Quality (TCEQ) - conducted between 2003-2006 in the Houston Brazoria counties - also focused on flares and it also highlighted the need to seriously address industrial flare operations. Over-steaming became and remains an area of intense interest. A further technical study commissioned by the TCEQ in 2008 further provided quantitative analysis on the significance of over-steaming practices and clearly demonstrated that air or steam-assisted flares can easily be "over assisted" resulting in very low combustion efficiencies. The TCEQ study clearly demonstrated the complex and inter-related impacts of multiple variables on flare performance, indicating that the management of flare combustion efficiency is a highly dynamic process. The EPA's response to this growing body of information has been enforcement first, with no regulatory improvement to date. Flare Consent Decrees Consent Decrees (CD) result from formalized enforcement negotiations between the EPA/state environmental protection agencies, the Department of Justice, and corporations/individual facilities. CDs can include multiple regulatory areas and can result in both civil penalties and enhanced environmental requirements. The EPA began to negotiate Global CDs with the refining industry following the introduction of the National Refinery Initiative in 2000 focusing on air emissions that included fugitive emissions, NSR/PSD permitting, benzene waste operations, ODS, EPCRA reporting, SO2 emissions and acid gas flaring, and a variety of regulatory issues. Since March 2000, the Agency has entered into approximately 31 settlements P a g e | 2 General Overview of Past and Recent EPA Flare Enforcement Actions with U.S. companies that represent over 90 percent of the Nation's petroleum refining capacity. These Global Consent Decree settlements cover 107 refineries in 32 states and territories and most do reference some flaring issues - mostly SO2 and acid gas events over 500 pounds per day, but these do not address the over-steaming issue. The EPA has negotiated Flare CDs that include enhanced requirements above and beyond the scope found within the federal flare regulations promulgated in 40 CFR 60.18 or 63.11 with the following four companies or individual facilities (date of entry provided in parentheses):  Equistar Chemicals - affects 7 olefins manufacturing plants in 4 states (July 2007);  INEOS Lanxess - affects the Addyston, OH butadiene plant (July 2009);  Marathon Petroleum Corporation - affects 6 refineries in 6 states (April 2012); and,  British Petroleum - affects the Whiting, IN petroleum refinery (May 2012). It should be stressed that the CDs noted above contain more than just flare related settlements. The following table provides a comparison of the respective CDs' flare provisions which demonstrates the progression and continuing evolution of thought regarding flare monitoring, performance, and operation within the agency. P a g e | 3 General Overview of Past and Recent EPA Flare Enforcement Actions Table 1 - Summary and Comparison of Requirements from Four Flare CDs Classification 1 Classification 2 Equistar 2007 INEOS Lanxess 2009 Marathon Petroleum 2012 BP Products Whiting 2012 Operating Parameters Steam to Vent Gas Ratio (mass) 0.91:1 Auto 3:1 Max 3:1 Max 3.6:1 Max Steam to Vent Gas Ratio (volume) 2.7:1 Max NHVFG (in the flare gas, at the tip) > 300 Btu/scf > 385 Btu/scf (or >200), set by P-FTIR >300 Btu/scf (variable NHV) > 300 Btu/scf (variable NHV) Flaring Limitations X X (SSM included) (by flare, by refinery) (2.1MMscfd)) Momentum Flux Ratio - MFR All > 0.0030; > 0.0030 (minimum) Detroit >0.0005 No Visible Emissions X X (SSM excepted) Exit Velocity (<60 fps) X X (with exceptions) X (with exceptions) DCS Logic Supplemental Fuel Addition X X X Steam Addition X X X Monitoring (Most all to be in 5 minute block averages, GCs every 15 minutes.) Vent Gas FR (includes T and P) X X X (dual channel) X Enhanced Vent Gas FR Monitoring X X 1,3 Butadiene FR X Steam FR X X X Enhanced Steam FR Monitoring X X Supplemental Fuel FR X Pilot Gas FR X Optional Optional Vent Gas MW X X X Vent Gas Composition (by GC) X X X Upstream of any Sweep or Purge Gas (thru C5+) (thru C5+) Wind Speed X X (Met station "high as possible") Video Camera X (if >5% in manual mode; 4 fpm) X (if >5% in manual mode; 4 fpm) P a g e | 4 General Overview of Past and Recent EPA Flare Enforcement Actions Classification 1 Classification 2 Equistar 2007 INEOS Lanxess 2009 Marathon Petroleum 2012 BP Products Whiting 2012 Report Development Flare Monitoring SOP X X X Enhanced Flare Monitoring SOP X X Flare Minimization Plans X X Flaring Minimization Implement Specific Minimization Techniques Tailchase FGR Systems (4) Performance Test PFTIR Test X (to determine NHVFG) X (at Garysville GF) X (2 flares) Training Operators & Supervisors X X Facility Survey Survey all sweep/purge gas meters X Survey/Minimize Sweep/Purge Gases X Survey HP HC PRVs X Root Cause Analysis Perform RCA on Reportable Flaring Events X X X P a g e | 5 General Overview of Past and Recent EPA Flare Enforcement Actions Comparison of Flare CDs A comparison of the flare related provisions in each of these four CDs demonstrates that the EPA has significantly increased its awareness of the over-steaming issue, and overall flare operations, and fully understands the technical ability of facilities to control flare performance with a high degree of control. Truly, the EPA has assimilated the available information relevant to flaring and combustion efficiency in order to negotiate more stringent requirements in the more recent CDs. EPA has aggressively used its enforcement "powers" prior to the signing of a CD to force companies into extensive and costly research programs. Very few provisions that are included in the INEOS Lanxess, MPC, and BP Whiting CDs are present in the 2007 Equistar CD, which was negotiated only two years prior to the INEOS CD and five years prior to the MPC and BP Whiting CDs. For instance, the Equistar CD does not include any provision or requirement to actively control the S/VG ratio or to monitor the majority of the flows that are required to be monitored in the later CDs. In fact the only operating parameter that is required to be monitored and adjusted at the Equistar facilities is the NHVcz. Monitoring that is explicitly required by Equistar is a continuous compositional analysis of the flare gas using gas chromatographs (GC) that were already installed at the facility prior to the CD negotiation. This review demonstrates that the provisions in the recent MPC and BP Whiting Flare CDs are much more detailed and robust than those found in the INEOS Lanxess CD. For example, the more recent 2012 CDs include increased monitoring frequencies, monitoring of additional parameters, and a greater degree of control of the flare operation by utilizing the increased amount of data for additional calculations. The two more recent CDs also include greater detail on the methods used to monitor the flow rates, including the specification that all flow rates be corrected for temperature and pressure. It clear from this quick analysis of these four flare CDs issued over the past five years that EPA is moving forward with more strict and more stringent requirements for flare performance. EPA is using its enforcement powers to move this shift in technological evolution, rather than its rule making processes. The published CD provisions are causing a significant shift of perspective with regards to industrial flare operations. Most refinery and some chemical facilities are beginning to operate their flares conceptually as a "process unit" instead of just a combustion device for excess vent gases in the system. While there are only a few facilities that are currently directly affected by these flare CD provisions, it is fully anticipated that the EPA will seek to include similar requirements, if not all of these requirements (or more new ideas like limitations on the Lower Flammability Limit), in future rule-making. This potential is causing all industrial flare operators to reexamine their current approach to efficient and compliant flare combustion. The Office of Air Quality Planning and Standards has recently prepared a report for peer review by the Flare Review Panel entitled "Parameters for Properly Designed and Operated Flares" (April 2012) that may prove to be the basis or foundation for a highly-anticipated Uniform Standard for Flares (40 CFR 65 rules) that may be proposed in the near future. P a g e | 6 General Overview of Past and Recent EPA Flare Enforcement Actions A Review of Recent Flare Assessment Activities Conducted by Sage Environmental Gap Assessment Tools While multiple approaches to attain compliance with the enhanced flare requirements may exist, the following list of gap assessment tools and compliance techniques has been compiled to aid facilities and flare operators to determine the current status of their flare systems. The primary goal of our current efforts focuses on flare performance optimization (awareness and monitoring improvements) and detailed process assessments to help minimize flaring, which can be achieved by pursuing the following three strategies, which are presented in no specific order:  Maximize gas recovery from the flare header,  Minimize contributions to the flare header, and  Optimize flare combustion and destruction efficiencies. The following gap assessment tools help to systematically identify opportunities in which these three strategies may be applied, there by achieving the primary goal of flare minimization. 1. Flare System Assessment. The flare system assessment helps to identify the current configuration of the flare compliance program by focusing on the following seven issues:  Flare Design,  Flare Operation,  Flare Emissions,  Monitoring,  Flare Minimization,  Recordkeeping,  Reporting, and  Testing. By completing the assessment, a facility can identify the key issues that need to be improved and/or upgraded and thereby prioritize action items to achieve increased gas recovery, minimized flare header contributions, and optimized flare performance. 2. Flare Mapping. Flare mapping systematically identifies each contribution to the flare header. This activity is necessary in order to truly minimize all contributions to the flare header as it is necessary to identify all contributions to the flare header prior to minimizing those contributions; it is difficult to minimize unknown flows and contributions. Mapping requires a detailed documentation of each contribution. Although the process to develop the flare map may be complex and time consuming, the result is more than a simple drawing consisting of the flare header and all contributions to the header. Rather, each contribution is documented (PRV, Control valve, sample port, etc…) and the reason or purpose for its use is discussed. The following figure is an example of a map of all contributions to the flare header: P a g e | 7 General Overview of Past and Recent EPA Flare Enforcement Actions Figure 1 - Example of a Map of all Contributions to the Flare Header P a g e | 8 General Overview of Past and Recent EPA Flare Enforcement Actions 3. PRV Valve Surveys. A valve leakage survey is commonly an acoustic measurement of potential in-line valve leakage of all valves that are typically closed during normal operations such as pressure relief valves connected to the flare header. Some have equated it to LDAR for flares headers (reminiscent of similar ambient surveys of valve fugitive emissions). The objective is not to specifically quantify the leak rate, but to identify and repair, when feasible, those significant in-line leaking valves in order to minimize the volume of gases that are directed to the flare header. Many valve surveys are conducted with monitors that utilize acoustic techniques, such as the new Mistras VPAC™II. This valve leakage survey is required currently of all refinery flare systems in California, and is a requirement of the BP Whiting Flare CD. 4. Flare Documentation, Minimization, and Analysis Tool (DMAT). The DMAT is a spreadsheet tool developed by Sage that provides a framework to systematically assess and qualitatively categorize and rank the various process related contributions to the flare header. Once the flare mapping is completed, each contribution to the flare header can be entered into the DMAT along with its typical frequency of contribution, and its typical flow rate, which enables the facility to easily and qualitatively identify the greatest contributions to the header. An example of a completed DMAT is provided in the following figure. P a g e | 9 General Overview of Past and Recent EPA Flare Enforcement Actions Figure 2 - Example of the Flare Documentation, Minimization, and Analysis Tool (DMAT) P a g e | 10 General Overview of Past and Recent EPA Flare Enforcement Actions Compliance Techniques 1. Flare Supervisor & Operator Training. Training may be necessary to instruct flare supervisors and operators regarding proper flare performance. Historically, flare operators have been instructed to operate flares in such a manner that there are no visible emissions. However, this has led most operators to over-steam the flare in order to ensure that visible emissions never occur at the flare. Training content may include the following topics:  Incipient Smoke Point (ISP),  Combustion efficiency - why "invisible flares" are potentially bad,  Need for and transition to automatic controls, and  Applicable regulations. 2. Monitoring Program Improvement and Development of Flare Monitoring Systems Protocol. The improvements to monitoring programs and the development of the monitoring protocol and plan document may be performed simultaneously in order to improve the efficiency of completing these items. Facilities developing a robust Monitoring Plan may consider including the following three sections:  all information, drawings, and diagrams regarding the flares and the flare monitoring equipment,  detailed descriptions of each monitoring instrument and piece of monitoring equipment, and  a description of the monitoring methods and calculations to comply with the monitoring requirements. 3. DCS Logic Development. DCS control and logic should incorporate and utilize all of the available data that is generated by the monitoring program to provide the flare system with a decision point on which to control the operating parameters and optimize the flare performance. Naturally, the quest to derive a single over-riding control parameter is elusive, and thus, this effort may require multiple assessment and programming considerations. At a minimum, a weekly internal report and operational review of the recent performance data, coupled with operating notes, will be critical in evaluating the site specific characteristics and actions necessary to optimize individual flare performance. 4. Flare Gas Minimization Plan. A written minimization plan is the culmination of all information gathered above by the various gap assessment tools and other compliance techniques, documenting the effort and ability of the facility to realize true minimization of flaring. Typical minimization plans lay out individual steps and projects that will be completed to achieve flaring minimization, focusing on the current contributions that possess the greatest opportunity to be minimized. This plan typically focuses on minimization during both normal operations and during maintenance, including major plant startup and shutdown activities (malfunctions and emergencies excepted). To be effective, these plans should be routinely reviewed and improved. Naturally, P a g e | 11 General Overview of Past and Recent EPA Flare Enforcement Actions considerable management involvement and commitment is necessary to continuously improve performance and realize flaring minimization. For further discussions or inquiries, please feel free to contact the presenter at: Troy M. Boley, Ph.D. Vice President and General Manager Sage Environmental Consulting, LP Kennesaw, Georgia Mobile [phone] Email [e-mail]