| OCR Text |
Show C O N T R O L L I N G S T E A M -‐ A S S I S T E D F L A R E O P E R A-T I O N S I N L I G H T O F F E D E R A L R E G U L A T I O N S Philip J. Smith, Jeremy N. Thornock, Sean T. Smith, Michal Hradisky CRSim Inc., Draper, Utah A B S T R A C T The difficulty of measuring combustion efficiency from flares has lead the EPA to promulgate procedural-based regulations rather than regulations based on achieving fuel emissions or combustion efficiency performance standards. These procedure-based regulations include the requirement to have an exit velocity below 18.3 m/s and a heat-ing value above 11.2 MJ/scmi. In this paper we explore the efficacy of this regulation by using a high-performance computing (HPC), large-eddy simulation (LES) code previ-ously validated to predict combustion efficiencies within 0.3% to examine a prototypical steam-assisted flare operating with fuels being fed under a range of exit velocities and heating values. The resulting combustion efficiency and unburned hydrocarbon emis-sions are analyzed in light of the federal regulation. I N T R O D U C T I O N The U.S. Environmental Protection Agency (EPA) has an ongoing priority for evaluating and bringing flare operations into regulatory compliance. Specifically, under the Air Toxics National Enforcement Initiative,ii the EPA is encouraging companies to report and correct any flare operations that the company may believe to result in poor combus-tion efficiencyiii. This reporting and regulatory compliance is based on the assumption that the operational procedures outlined by the EPA will in fact result in improved com-bustion efficiency over operations that are not aligned with the existing regulations. Maximizing combustion efficiency is the the objective for good flare operations. Moni-toring combustion efficiency is difficult, so the EPA has promulgated procedural based regulations. It is the purpose of this paper to study one of these procedural based regula-tions; namely, NSPS 40 CFR Part 60, Subpart A (§60.18), to understand the impact of the procedure on combustion efficiency. Current high performance computing (hpc) technology has made it possible to compute combustion efficiency from flares with a high degree of accuracyiv. These techniques employ large eddy simulations (LES) that couple the simulation of turbulent mixing through dynamic coherent structures with the flame chemistry and heat transfer to pre-dict the dynamic behavior of the flare. The combustion efficiency is computed locally version: Friday, August 30, 2013 i NSPS 40 CFR Part 60, Subpart A (§60.18): "Flares shall be used only with the net heating value of the gas being combusted being 11.2 MJ/scm (300 Btu/scf) or greater if the flare is steam-assisted or air-assisted; or with the net heating value of the gas being combusted being 7.45 MJ/scm (200 Btu/scf) or greater if the flare is nonassisted ... and... operated with an exit velocity ... less than 18.3 m/sec (60 ft/sec)." ii This initiative along with a map of locations for enforcement actions taken in 2012 are described on the EPA web page: http://www.epa.gov/compliance/data/planning/initiatives/2011airtoxics.html iii EPA Enforcement Alert EPA 325-F-012-002 iv Jatale, Anchal ; Smith, Philip J.; Thornock, Jeremy ; Smith, Sean, "Validation of Flare Combustion Effi-ciency Simulations," American Flame Research Committee, Salt Lake City, 2012. and globally to within 0.3%. The overall combustion efficiency is governed by the local mixing, reaction and extinction. In this paper we study the effect on combustion efficiency as computed by the LES hpc simulations of using the EPA procedures for one steam-assisted flare. T C E Q S T E A M -‐ A S S I S T E D F L A R E T E S T Much of the research that the EPA regulations have been built on, were for well con-trolled laboratory experiments. Questions have been raised as to the applicability of these small-scale experiments to full scale industrial flares. Combustion efficiency is dif-ficult to measure in full-scale elevated flares. A variety of methods have been proposed. The Texas Commission on Environmental Quality (TCEQ) contracted with The Univer-sity of Texas at Austin to conduct the Comprehensive Flare Studyv project to study full-scale operations and measurement methods. The objective of this project was to deter-mine the effect of vent gas flow rate turndown and steam assist rates on full-scale indus-trial flare combustion efficiency as measured with a variety of measurement techniques. Specifically the field tests for the steam-assisted flare were conducted using John Zink Model EE-QSC-36" Flare Tip with three EEP-503 pilots having a maximum capacities of 937,000 lb/hr. The test results have been reported elsewherevi. The steam-assisted flare was simulatedvii with the most complete geometry available as seen in Figure 1 and details of the resolution obtained for the simulation are shown in Figures 2 & 3. Figure 1: CAD image of the Zink steamed flare head for the TCEQ steam-assisted flare simulation studied in this paper. E v a l u a t i n g S t e am-‐ A s s i s t e d F l a r e o p e r a t i o n s page 2 of 5 v PGA No. 582-8-86245-fy09-04, Tracking Number 2008-81, TCEQ, 2009 and TCEQ Grant No. 582-10-94300. vi Texas Commission on Environmental Quality PGA No. 582-8-862-45-FY09-04 Tracking No. 2008-81, TCEQ 2010 Flare Study Final Report, prepared by David T. Allen, Ph.D. Principal Investigator and Vincent M. Tor-res Project Manager. vii The base case for the simulations of this study was that reported as S4.1 in the TCEQ report. T H E L E S A N A L Y S I S The Simulator: The disparate scales ranging from the steam injection ports to the far-field flare plume behavior requires integration of two computa-tional approaches. We perform the flare plume computations with our in-house LES code (Arches). Arches is capable of scaling to large numbers of processors (demonstrated up to 300,000 processors) with a long history of combustion modeling development. We perform the flow calculation in and around the complex geometry of the flare tip and steam injectors with Star-CCM+, a commercial code with demonstrated scaling up to ~500 processors. While each code differed in how they treated the LES combustion algorithm, they did share a common goal of using increased explicit representation of the turbulent scales on the LES mesh to tie information from the resolved scale to the sub-grid scale. The mesh resolution changes as we pass information from one code to the other. One example of the information passed between codes at this handoff plane is shown in Figure 4. Figure 2: Close up of the computational mesh and geometry for the upper steam nozzles for the Zink steam-assisted flare studied herein. Figure 3: Detail of one injection port of one arm of the upper steam nozzles shown in Figs. 1&2. Note that there are over one hundred mesh points around the circumference of each port. E v a l u a t i n g S t e am-‐ A s s i s t e d F l a r e o p e r a t i o n s page 3 of 5 One of the flare simulations is shown in the volume rendered frame from an animation of the dynamic flare simulation in Figure 5. The Simulations: We have performed a series of four simulations to study the effect of heating value on the per-formance of the TCEQ steam-assisted flare. The base case was that of test S4.1 from the aforementioned TCEQ study. In this case the fuel was mostly pro-pylene and nitrogen with an heating value of 350 BTU/scf and an EPA exit velocity of 2 fps. Figure 6 shows a snap-shot in time of the volume rendered temperature field from the base case simulation. Three variations on this base case were studied. Two cases were studied with lower heating values (175 BTU/scf and 88 BTU/scf) with all other operating parameters fixed to be that of the base case; including exit veloc-ity. Both of these two cases are below the EPA allowable heating value. One additional case was studied with no steam but the same fuel as the base case (350 BTU/scf) which is within that allowed by the EPA. Figure 4: Representation of the velocity vectors at the handoff plane between the Star CCM+ and Arches for the Zink steam-assisted flare. Figure 5: Volume rendered image of a frame from an animation sequence of the Zink flare of the TCEQ steam-assisted flare test. E v a l u a t i n g S t e am-‐ A s s i s t e d F l a r e o p e r a t i o n s page 4 of 5 The evaluation showing animations of each of these four cases can be seen at http://crsim.com C O N C L U L S I O N S We have studied the performance of low flow-rate, low heating-value vent gases under varying wind conditions. The study was conducted by computing combustion efficiency with validated hpc LES com-puter codes that have been consistent with combustion efficiency measure-ments in wind tunnels to +/- 0.3%. This study explored the efficacy of the EPA operational regulations regarding com-bustion efficiency. Figure 6: A frame from an animation sequence showing the volume rendered temperature field in the base case simulation of the Zink flare of the TCEQ steam-assisted flare test. E v a l u a t i n g S t e am-‐ A s s i s t e d F l a r e o p e r a t i o n s page 5 of 5 |
