Transient Ignition of Multi-Tip Ground Flares

Paper from the AFRC 2018 conference titled Transient Ignition of Multi-Tip Ground Flares

Multi-Point Ground Flares (MPGF) represent a special class of flares capable of processing significant quantities of flare gas in flare fields which includes hundreds of burners surrounded by a complex wind fence that limits radiation to surrounding equipment and improves safety for plant personnel. A detailed computational fluid dynamics (CFD) model of an MPGF has been developed using a proprietary flare modeling tool called C3d. This tool has been used to simulate the ignition phenomena for several MPGFs for flare gas flow rates between hundreds to 1000 tons or more per hour (TPH).; Simulation results have been directly compared to operating test data for an elevated multipoint methane flare for calibration. Results demonstrate the ability of C3d to replicate the measured flame spread rate and reproduce the measured pressure wave generated during the ignition event. Based on this validation, the tool has been used to conduct over sixty separate simulations to investigate the ignition behavior for this multi-point elevated flare. Results from these simulations clearly show the critical effect of ignition delay on the magnitude of the pressure wave generated on ignition. The main conclusion drawn from this analysis is that the ignition system's reliability to quickly ignite the flare gas above the flare tip is critical to safe operation. Predictions show that a 0.6 second ignition delay results in a significant pressure wave generated during flare ignition. Simulations at maximum flow rate (1350 TPH) exhibit explosive tendencies with pressure waves greater than one atmosphere. This confirms the conclusion that the flare must be operated with a continuous pilot to avoid and type of ignition delay. These results underscore the importance of the API recommended practice of continuous pilot operation for all large scale gas flares.; The application of this tool to MPGF has demonstrated its ability to cause ignition in adjacent unignited rows due to wind effects. The tool is capable of predicting pressure wave events in these unignited row scenarios. Such effects can be important when pilots may be extinguished and combustible hydrocarbons are released to the atmosphere in a well-mixed and highly combustible state. Such gas clouds could be ignited remotely and create large pressure waves capable of damaging surrounding plant equipment and injuring plant personnel.

American Flame Research Committees - AFRC 2018 Industrial Combustion Symposium University of Utah, Salt Lake City, Utah September 17-19, 2018 Transient Ignition of Multi-Tip Ground Flares Joseph D. Smith, Ph.D., Laufer Endowed Energy Chair Missouri University of Science and Technology, Rolla, Missouri, USA Robert E. Jackson, Zachary P. Smith and Ahti Suo-Anttila, Ph.D. Elevated Analytics, Inc., Catoosa, Oklahoma, USA Doug Allen and Scot Smith Zeeco Inc. Broken Arrow, Oklahoma, USA ABSTRACT Multi-Point Ground Flares (MPGF) represent a special class of flares capable of processing significant quantities of flare gas in flare fields which includes hundreds of burners surrounded by a complex wind fence that limits radiation to surrounding equipment and improves safety for plant personnel. A detailed computational fluid dynamics (CFD) model of an MPGF has been developed using a proprietary flare modeling tool called C3d. This tool has been used to simulate the ignition phenomena for several MPGFs for flare gas flow rates between hundreds to 1000 tons or more per hour (TPH). Simulation results have been directly compared to operating test data for an elevated multipoint methane flare for calibration. Results demonstrate the ability of C3d to replicate the measured flame spread rate and reproduce the measured pressure wave generated during the ignition event. Based on this validation, the tool has been used to conduct over sixty separate simulations to investigate the ignition behavior for this multi-point elevated flare. Results from these simulations clearly show the critical effect of ignition delay on the magnitude of the pressure wave generated on ignition. The main conclusion drawn from this analysis is that the ignition system's reliability to quickly ignite the flare gas above the flare tip is critical to safe operation. Predictions show that a 0.6 second ignition delay results in a significant pressure wave generated during flare ignition. Simulations at maximum flow rate (1350 TPH) exhibit explosive tendencies with pressure waves greater than one atmosphere. This confirms the conclusion that the flare must be operated with a continuous pilot to avoid and type of ignition delay. These results underscore the importance of the API recommended practice of continuous pilot operation for all large scale gas flares. The application of this tool to MPGF has demonstrated its ability to cause ignition in adjacent unignited rows due to wind effects. The tool is capable of predicting pressure wave events in these unignited row scenarios. Such effects can be important when pilots may be extinguished and combustible hydrocarbons are released to the atmosphere in a well-mixed and highly combustible state. Such gas clouds could be ignited remotely and create large pressure waves capable of damaging surrounding plant equipment and injuring plant personnel. Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency 1. Introduction Flares are an important safety device used during upset or sudden shutdown events in a refinery or chemical plant. They are used to combust relief and unusable flammable gases in an open atmosphere to prevent explosions that can lead to tragic events plants and surrounding facilities. Also, they are used to prevent atmospheric pollution from process gases by burning them to carbon dioxide and water vapor. The world bank has supported an international effort in many countries aimed at reducing the total amount of gas flared around the world [1]. However, since 2016 the global trend of gas flaring has increased each year due to increased global oil and gas production. In 2012, an estimated 145 billion cubic meters of gas was flared compared to 149 billion cubic meters in 2016 [2], [3]. Additionally, regulations for flaring operation and design (e.g., EPA 40 CFR 60.18 and API 521) have evolved to ensure improved flare performance [4, 5]. However, during normal flare operation, ignition can be delayed for unlit flares or extinguished pilot flames which lead to excessive releases of unburned hydrocarbons and black carbon (BC) into the atmosphere. During flare ignition, gas is sent from the process to the flare stack where the gas is ignited and burns above the flare tip with effluents released into the atmosphere. This during the ignition process, mixing between the flare gas and surrounding air is limited due to relatively low gas flow through the tip which leads to incomplete combustion where BC emissions are formed. This transient condition, caused by inefficient mixing, normally lasts approximately 10-30 seconds of the overall flare operation. This represents an uncontrolled source of large amounts of BC released into the environment. Black carbon has a detrimental influence on plant safety, on human health and the climate. For example, when black carbon deposits on reflective surfaces, e.g. snow, ice, and white surfaces, it darkens these surfaces and reduces the amount of reflected solar radiation and increases energy absorbed by these surfaces. This impacts the climate when these surfaces are covered by snow or ice since this increases melting and raises ground temperature. Also, BC influences cloud dynamics and atmospheric absorption of solar radiation. BC has also been shown to significantly affect human health by increasing respiratory illnesses, cancer, and congenital defects [6]. Approximately 10-30% of energy released during flare gas combustion radiates from BC to surrounding facilities and equipment. Therefore, flares are typically located away from equipment and buildings to avoid heat damage and to reduce the safety risk to workers. Accurate estimation of BC formation during flare ignition and operation and the associated thermal radiation from the flame is essential to minimize safety risk to plant personnel, to reduce damage to plant equipment and buildings, to minimize impact on human health and to reduce climate impact. Many studies on measuring and estimation of BC emissions from flaring have been carried out [7, 8, 9, 10, 11, 12, 13]. A common finding of these studies showed that BC emissions increased with decreasing flare combustion efficiency. McEwen et. al. measured the quantitative emission of soot in lab scale flares with inner diameters between 0.0127-0.0762 m and with a jet velocity between 0.1-2.2 m/sec for four and six component methane-based fuel mixtures [9]. They developed an empirical relationship between the flare gas heating value (HV, MJ/m3) and the soot emission factor (SEF, kg of soot/103 m3 fuel) to be: 𝑆𝐸𝐹 = 0.0578 ∗ 𝐻𝑉 − 2.09 Page 2 of 34 (1) Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Wang et. al., suggested a new reaction mechanism including 50 components that he used to estimate soot emission by predicting key soot precursors including C2H2, C2H4, and C6H6 [13]. They showed a CFD simulation of an air-assisted flare using this mechanism provided an improved prediction of the amount of soot emission from the flare. Several investigations have also been done to analyze the amount of thermal radiation emissions from industrial scale flare flames [14, 15, 16, 17, 18, 19]. Smith, et. al. evaluated the effect multiple flares on a neighboring multi-point ground flare (MPGF) system at maximum flare gas flow rates using a Large-Eddy Simulation (LES) based CFD analysis [17]. Smith et. al. also studied the effects of a flare plume produced by a MPGF on surrounding equipment, facilities and workers at maximum flow rates using an LES based CFD analysis [18]. Recently, a new modeling approach to predict heat radiation from gas flames was introduced by Miller [19]. However, this model was developed to study H2 and Syngas fires. Analytical quantification of an Multi-Point Elevated Flare (MPEF) operating at full-rate is very difficult due to the size of the combined flame and the associated high radiation flux from the flame operating high above the ground. These issues also make it impossible to measure emissions generated by an MPEF. The API 521 flaring guidelines include recommended design criteria for a flaring system but mostly focus on utility flares and elevated air and steam assisted flares with no consideration for MPGF or MPEF systems. Smith et.al. [14] and Smoot et.al. [20] point out that additional performance metrics are needed specifically for these flares to design and quantify their performance. To the author's knowledge, none of these studies mentioned above regarding BC emissions from flares have specifically considered the flare ignition process. Therefore, the objective of this work has been to estimate the amount of BC released during the ignition process of flare systems specifically. Another focus for this work was to study BC generated by flare flames and the associated thermal radiation to surrounding equipment and structures using an LES based CFD technique described below. 1.1. Foundational Work in Determining Flare Performance Combustion efficiency (CE) and Destruction Removal Efficiency (DRE) are accepted indicators of flare performance. 𝐶𝑂2(𝑝𝑙𝑢𝑚𝑒) 𝐶𝐸(%) = (𝐶𝑂2(𝑝𝑙𝑢𝑚𝑒)+𝐶𝑂(𝑝𝑙𝑢𝑚𝑒)+∑ ℎ𝑦𝑑𝑟𝑜𝑐𝑎𝑟𝑏𝑜𝑛 (𝑝𝑙𝑢𝑚𝑒)) 𝑥100 (2) where: CE(%) = combustion efficiency (%) CO2 (plume) = carbon dioxide concentration (ppmv) in plume after combustion CO (plume) = carbon monoxide concentration (ppmv) in plume after combustion ∑ ℎ𝑦𝑑𝑟𝑜𝑐𝑎𝑟𝑏𝑜𝑛 (𝑝𝑙𝑢𝑚𝑒) = unburned hydrocarbon concentration in plume after combustion multiplied by number of carbon atoms in hydrocarbon (ppmv) 𝐷𝑅𝐸(%) = (1 − 𝑋𝑝𝑙𝑢𝑚𝑒 𝑋𝑖𝑛 )𝑥100 where: DRE(%) = destruction and removal efficiency (%) 𝑋𝑝𝑙𝑢𝑚𝑒 = species mass flow X found in flare plume after combustion 𝑋𝑖𝑛 = species X mass flow in flare gas entering flare Page 3 of 34 (3) Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency When a flare's CE is equal to or greater than 96.5%, flare performance is judged as acceptable [21]. Previous studies have investigated and quantified the impacts of various operating parameters on flare performance [22, 23, 24, 25, 26, 27, 28, 29, 30]. The foundational work to determine flare performance were conducted at the John Zink flare testing facility located in Tulsa, OK. This work analyzed flame stability and flare CE and DRE efficiency for various vent gas compositions and heat content, flare tip diameter, flare gas flow rate and tip velocity, ambient wind speed, and steam or air to flare gas mass ratios. McDaniel [26] studied the CE and DRE for both steam assisted and air assisted flares under a wide range of operating conditions including flare gas flow rate, flare gas heat content and steam to gas mass ratio. The steam flare effective diameter was 0.149 m while the air flare used two spider burner tips with holes having an effective exit area of 0.00342 m2 for high heating value gas and 0.0073 m2 for low heating vent gas. Mixtures of propylene diluted with nitrogen were used as the flare gas. He found CE's decreased for steam to flare gas ratio greater than four and for high tip velocities and heating values below 11,178 KJ/scm (300 Btu/ft3). The main conclusion from his work was that gas flares have destruction efficiencies greater than 98% when properly operated. Pohl et al. [23] investigated the effect of flare gas tip velocity, steam to flare gas mass ratio (S/FG), and flare tip size on CE. He tested pipe flares with tip diameters of 0.0762, 0.1524, and 0.3048 m in the absence of wind. He also tested steam flares with varying steam flow rate ranging up to 1 kg steam to 1 kg of flare gas for tip velocities ranging between 0.061and 130 m/sec. He tested mixtures of propane-nitrogen with heating values ranging between 10,060 - 87,559 kJ/scm, (270 - 2,350 Btu/scf) and methane (natural gas) as flare gas. His findings confirmed that assisted flares with pilots were more stable than those without pilots. He also found that flame stability (defined as when tip velocity exceeds known flame velocity) was correlated with flare gas heat content and flare tip velocity. In another study, Pohl [24] studied the impact of flare tip design with several flare gas compositions on CE. In this study, he tested a Coanda flare tip (provided by Flaregas Corporation) with steam injected as the entraining fluid through a 0.3048m diameter opening. He also tested an air-assisted flare with a tip diameter of 0.0381m with flare tip velocities between 0.061 to 0.251 m/sec. He confirmed flare tip design affected flame stability and showed that CE and flame stability were correlated to the air to flare gas mass ratio and net heating value for the air assisted flare. Castiñeira and Edgar [22] studied the impacts of steam and air to fuel ratios on the CE of flares. They used a 2-D simulation to conduct their study. CFD simulations showed that inefficient hydrocarbon combustion was predicted to occur at high air to fuel and steam to fuel mass ratios. In a follow-on study, they studied the effect of flare gas tip velocity to crosswind velocity momentum ratio (i.e., defined as momentum flux ratio by Seebold [31] ) on flare performance [32]. Their simulations of natural gas flames in a closed-loop wind tunnel showed that CE declined at low momentum flux ratio (high cross wind velocity). The Texas Commission on Environmental Quality (TCEQ) in cooperation with the University of Texas at Austin investigated the effects of low flare gas flow rates (0.1 % and 0.25 of total flare firing capacity), heat content of 13041 kJ/m3 (350 Btu/ ft3), 22355 (600 Btu/ ft3), and 80069 kJ/m3 (2,150 Btu/ft3) together with three steam-assist flow rates (0.0, 425, and 1062 kg/hr) and two airassist flow rates (163 and 425 kg/hr) on DRE and CE [27]. These tests, conducted at the John Zink Flare Test Facility in Tulsa, OK, used both extractive sampling and remote PFTIR and AFTIR Page 4 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency (Passive and Active Fourier Transform Infrared) techniques to sample the flare plume at 1 sample per second (1 Hz)> The main objectives of this work included: • • • • Investigate impact of flare gas flow on CE and DRE, Investigate impact of assist media (steam and air) flow rate on CE and DRE at low flare gas flows, Assess effectiveness of remote sensing systems to quantify flare emissions contained in the flare plume, and Assess whether flares operated within EPA regulations outlined in 40CFR60.18 achieved a DRE ≥ 98%. These tests were conducted with a steam assisted flare with an effective diameter of 0.9144 m and an air assisted flare with an effective diameter of 0.6096 m. Tests considered flare gas compositions made up of mixtures of Tulsa natural gas, propylene, and nitrogen. Results illustrate how complex this issue is. When propylene was flared at low flow rates with steam, the flare had a DRE ≥ 98% when the steam to flare gas mass ratio (S/FG) was ≤ 0.29 1 for a flare gas heating value of 300 Btu/scf. When the gas heating value was increased to 600 Btu/ft 3 the steam flare achieved a DRE ≥ 98% for a S/FG ≤ 0.82 which showed the impact flare gas heating value has on flare performance. When the flare gas heating value was held constant at 350 Btu/ft3 for a low flare gas flow rate, flare performance varied (i.e., DRE changed) depending on where the steam was injected. When steam flow rate and how it was injected were held constant, the DRE increased as flare gas flow rate was increased. For air assisted flares, when flare gas heating value and flow rate were held constant at 359 lb/hr and 937 lb/hr, the DRE remained above 98% but decreased when the amount of assist air increased above 6 times the stoichiometric requirement for both propylene and propane mixtures. The TCEQ also elevated several remote sensing technologies including PFTIR, AFTIR and IR (using the LSI FLIR GasFindIR camera). FLIR images indicated unburnt hydrocarbons in the flare plume but could not provide quantitative measurements of flare performance. Measured CE values using AFTIR and PFTIR showed more than a 2% difference compared to extractive sampling. This difference is large considering the required DRE must be >98% so when the FTIR systems suggested the flare was meeting the standard it could actually be operating below the required DRE level. 1.2. Factors Important in Flare Performance As shown by the previous work described above, flare performance quantified by Combustion efficiency (CE) and Destruction Removal Efficiency (DRE) is affected by many parameters. Interested readers are encouraged to review the report prepared by the EPA that describe each of the key parameters in detail [33] with only the main factors discussed in the following sections. 1.2.1. Exit velocity (Momentum ratio) The Momentum Ratio (MR), or Momentum Flux ratio discussed by Seebold [31], is the ratio of flare gas momentum and crosswind momentum. This parameter assesses the impact of ambient wind on flare performance. Flame deflection and poor mixing results when wind momentum dominates (i.e., the Momentum Ratio is low) which occurs during low flare gas flow [34]. If the 1 The EPA suggests the steam to flare gas ratio be set to the flare manufacturer recommended value. The Steam to Flare Gas ratio required to achieve a DRE>98% was less than what was recommended by the flare vendor. Page 5 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency MR < 10%, then the flame can stabilize on the downwind side of the flare stack [35]. As cross wind velocity increases, flame height decreases and the flame may stabilize inside the flare tip causing heat damage or it may detach from the stack (see Figure 1). Detached flames may also occur at high exit velocity which Pohl showed impacts flame stability [23]. Castiñeira and Edgar [22] also showed when unstable combustion occurs flare performance is degraded. However, other factors also impact flare performance including flare gas heating value, amount of assist media provided and how it is provided, and flare tip design. Pohl originally showed [23] and Evans [36] reaffirmed based on work measuring industrial gas flare efficiency that though tip exit velocity is important to flare performance, it does not control (see Figure 2 and Figure 3). Evans also pointed out that based on performance collected from large industrial flares that smoking does not imply low CE [36]. Seebold [37] supported this point by comparing the average measured CE for 45 smoking flares (99.2%) to that for 55 non-smoking flares (98.1%) as shown in Figure 4. Figure 1 - Flare flame shapes in high cross winds [31] Page 6 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 2 - Flame stability showing high flare efficiency at very tip high velocity [23], [36] Page 7 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 3 - Tip velocity vs heating value per EPA regulations compared to Pohl's flame stability chart [36] compared to measured flare efficiency [38] for tip velocity vs heating value Page 8 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 4 - Comparing measured CE for smoking and non-smoking flares [37] 1.2.2. Flare Gas Heat Content (< Minimum Stability Criteria) Low energy flare gas supplied to flare combustion zone has been shown to result in lower CE mostly because of lower flame temperatures. Lower flame temperatures reduce the reaction rate and make the flame more susceptible to flame shearing and quenching. Pohl discussed this originally in terms of flame stability (see Figure 2) verses flare gas heating value and tip exit velocity [23]. According to Castiñeira and Edgar [22, 32] and Smoot et.al., [20], flares operate with low CE when the flare gas has a heating value below 200 Btu/scf. Also, as discussed earlier and shown in Figure 2, Pohl showed saturated hydrocarbons with high heat content (>1000 Btu/scf) burned with a stable flame even at very high exit velocities (> 400 ft/s). Assist media lowers the combustion gas heat content but it also increases the amount of ambient air with the flare gas which improves flare CE if the amount of assist media is below established limits (e.g., < 4 pounds steam per pound of flare gas with heat value > 200 Btu/scf; < 6 times stoichiometric air). Figure 5 shows how sensitive flare CE is to heat content for a mixture of propane and nitrogen without a pilot flame [23]. Practical flare operation following API guidelines [39] as practiced by companies such as Marathon Petroleum Company [40] documents the requirement to use higher levels of assist media to burn hydrocarbons with higher carbon to hydrogen ratios such as ethylene (0.4 to 0.5 lb steam/lb ethylene) and butadiene (0.9 to 1.0 lb steam/lb butadiene) without smoke. Page 9 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 5 - CE near lower limit of stability for mixture of C3H8/N2 without pilot flame [23]. 1.2.3. Assistant Medium Steam and Air are used as assist media in flares to suppress smoking. The purposes of using assist medium is to add momentum to flare vent gas, to enhance turbulent mixing between the flare gas and ambient air, and to induce more air into the flare combustion zone. According to Smoot et al., over steaming occurs when more than four pounds of steam per pound of flare gas is used [20]. They point out that even though using steam increases the amount of air in the combustion zone, over steaming dilutes and quenches (cools) the reaction zone and lowers the CE of the flare. The U.S. EPA Office of Air Quality Planning and Standards commissioned a comprehensive study that collected and analyzed key flare operating parameters shown to affect flare performance in previous studies [33]. Results presented in this study confirmed that assist air levels > 7 time the stoichiometric air requirement also resulted in degraded CE for most hydrocarbons [33]. Pohl et.al. [25] identified a Flame Stability region for various flare tip designs for various tip exit velocities versus flare gas heating values. As discussed earlier, flame stability is a complex issue affected by several parameters including flare gas heating value, turbulent mixing, and reaction kinetics. This fact is illustrated in Pohl's flame stability region were various tips exhibited greater stability than others. Page 10 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Table 1 - API 521 Table 11 Suggested Steam Rates [39] In summary, Seebold concludes flare CE are independent of \aerodynamic conditions and in all cases (except near the limit of stability), the combustion efficiencies are uniformly high - always >98% and usually >99% [37]. 2. Materials and Methods 2.1. Testing Several flare tests have been performed at the Zeeco Flare Test Facility in Broken Arrow, OK using the set up shown in Figure 7 and Figure 8. One set of tests was used to study soot emission and thermal radiation emissions from single tip and three tips flare. A view of typical flames recorded during these tests are shown in Figure 9. The purpose of these tests was to collect validation data to establish a CFD based flare model that could be used to analyze full scale multitip ground flare system designs. Page 11 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 6 - Flame Stability region of steam-injected and pressure assisted heads [25] Figure 7 - Zeeco multi-flare burner test setup [41] Page 12 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 8 - Test stand for single and multi-tip experiments In general, various flare tips have been developed for pipe (utility) flares, air/steam assist flares, and pressure assisted multi arm flare tips used in enclosed ground and multi-tip ground flare systems. A view of a multi arm burner is shown in Figure 10. As shown, the tip is positioned at the end of a vertical riser which connects to the main horizontal runner. One advantage of the multi-arm flare tip is that it uses gas pressure to create high velocity jets that improve local turbulent mixing of flaring gas and ambient air to ensure clean (smokeless) operation at high flow rates. These burners also minimize thermal radiation by burning radiating soot faster which results in shorter flames compared to normal flare tips. Finally, flames from these flare tips are more resistant to cross winds because the flare gas momentum is typically much greater than the cross wind momentum (e.g., momentum flux ratio is relatively large). These flare burner tips are normally fabricated from 310 stainless steel via investment casting to improve their heat resistance and to increase their operating life. To achieve safe and reliable operation, these flare burners are pressure tested in the shop and field tested for smokeless operation and radiation flux as illustrated in tests described below. Page 13 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 9. Flame shape for single tip and three tip flare test burning propylene [15]. Figure 10 - Multi arm flare burner tip (courtesy of Zeeco, Inc.) The flare tests described below, carried out at the Zeeco flare test facility in Broken Arrow, involved several flare gases including methane, propane, propylene ethylene and xylene. Thermal radiation measurements were taken during each test using radiometers located at various locations from the flare flame. Wind speed and direction, ambient temperature, barometric pressure, and relative humidity were measured using Zeeco's onsite weather station which records data at a sample rate of 2 Hertz, providing both wind gust information as well average conditions over the duration of the tests.. 2.1.1 Multi-Tip Ground Flare Testing A test was conducted to measure the flame height and flame shape for a multi-tip ground flare burner and to quantify the wind effect on flames from a single flare tip burning propane (see Figure 11). This work was done to help validate a CFD tool used to predict flare performance with results shown in Table 2 and originally reported by Smith, et.al. [42]. With no ambient wind, the flame shape shown was cylindrical and thin above the tip and approximately 13.8 to 15.3 m high. With Page 14 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency an 8 to 10 mph wind, the flame was deflected approximately 8 - 10° from the perpendicular and the height shortened to about 10.5 m with the shorter flame generating less thermal radiation due to flame aeration which consumed the radiating soot produced in the flame. Figure 11 - Wind effects on the flame shape of single flare tip with propane flow rate: measured at 1.4 in-wc @ 57 ˚F across orifice plate (7.3 psig tip pressure on 0.4572 meter (18 inch pipe)) [14] A second flare test was conducted with three flare burners located adjacent to each other firing ethylene in a no-wind condition as shown in Figure 12. The flare burner tips, similar to those shown in Figure 10, were located approximately 66" apart and fired at three relatively low tip pressures (2.8, 7.3, and 11.4 psig). For these conditions, the flames appeared to be approximately 11 m high and had a cylindrical shape above the flare tip. The flames remained distinct near the tip but merged approximately 3 m above the tips to form a large single flame. Using the image shown on the left, the CFD tool was applied to simulate the flame as shown on the right. These tests helped to validate the CFD tool which has been used to predict multi-tip ground flare performance as reported by Smith et.al. [42]. Another flare test was conducted using three flare burner tips (similar to those shown in Figure 10) to fire 5,465 kg/sec propylene with a tip pressure of 22.5 psig. This test was conducted at a average wind speed of 11.2 mph (gusting 9-13 mph) blowing from an average of 169 degrees from true north. In this work, Smith et.al. [42] reported the thermal radiation measurements recorded using radiometers located 22.9 m, 30.5 m, and 45.7 m (75 ft, 100 ft, and 150 ft) from the flare flames (see Table 3). The radiometers were positioned 1.5 m and 6.1 m (5 ft and 20 ft) above grade and located due east of the flare flame (see Figure 13 for test setup). As expected, thermal radiation flux was highest near the flame and was higher above the ground (20 ft) than near ground level (5ft). Page 15 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Table 2 - Measured Radiation Heat Flux from three flare tips burning ethylene taken at two locations Tip Size 3 3 3 3 3 3 4 4 4 4 4 4 Position (m) 15 15 15 30 30 30 15 15 15 30 30 30 Burner Pressure (psi) Measured Radiation (W/m2) 2.8 3344 7.3 4803 11.4 6192 2.8 671 7.3 1184 11.4 1532 2.8 6371 7.3 8192 11.4 9536 2.8 1513 7.3 2464 11.4 2747 Figure 12 - 3-tip flare test burning ethylene (conducted at Zeeco Flare Test Facility) with simulated flame shown overlaid on test image. Page 16 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 13 - Setup for radiation measurements from a multi-tip flare test Table 3 - Radiation from 3-tip flare test burning propylene Elevation (wind speed) 5 ft high (3-7mph wind) 20 ft high (3-7mph wind) Radiometer distance from flare Measured Flux (BTU/hr-ft2) Measured Flux (BTU/hr-ft2) 75 feet 171 205 100 feet 102 102 150 feet 34 34 Page 17 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 14 - 3-tip flare test burning propylene at Zeeco's Broken Arrow Test Facility [42] One complication in making flare radiation measurements is the need to account for radiation from the heated ground surrounding a flare to radiation meters. Since a flare emits significant radiation, the ground surrounding the flare heats up. The heated ground emits radiation that contributes to the overall radiation flux measured by the radiometer. In addition, the ground also reflects any radiation not absorbed. Using an emissivity and absorptivity set to 1.0 provides a reasonable approximation of both emitted and reflected radiation. The overall energy balance for a grey surface that both emits and reflects radiation is identical to that which absorbs all the incoming radiant heat and then re-radiates it to the surroundings at steady state. Thus, a radiation meter, which is assumed to be a grey or black surface, sees the same incoming radiation flux in both scenarios. If the ground reflects radiation spectrally, rather than diffusely, then some uncertainty is introduced. A second complication is related to the atmospheric transmissivity which as discussed by Hamins [45] is included for all radiation that passes through clear air. The Hamins model depends upon ambient temperature, source temperature, and relative humidity. It accounts for radiation absorption by water vapor and carbon dioxide in the radiation path from the source to the surface. 2.2. Combustion Reaction Mechanism Numerous reaction mechanisms have been suggested to properly simulate the complex turbulent combustion chemistry occurring in flare flames. Among these mechanisms, Smith et al. [16, 41] adopted the simplified four step chemical reaction mechanism first proposed by Said et.al. [47] to approximate the flare gas combustion. 1𝑘𝑔 𝐹 + (2.87 − 2.6 𝑆1 )𝑘𝑔 𝑂2 → 𝑆1 𝑘𝑔 𝐶 + (3.87 − 3.6 𝑆1 ) 𝑘𝑔 𝑃𝐶 + (50 − 32 𝑆1 ) 𝑀 (4) 1𝑘𝑔 𝐹 + 0.3 𝑀𝐽 → 𝑆2 𝑘𝑔 𝐶 + (1 − 𝑆2 ) 𝑘𝑔 𝐼𝑆 (5) 1𝑘𝑔 𝐶 + 2.6 𝑘𝑔 𝑂2 → 3.6 𝑘𝑔 𝐶𝑂2 + 32 𝑀𝐽 (6) 1kg IS +(2.87 - 2.6 𝑆2 )/(1-𝑆2 ) kg 𝑂2 → (3.87 − 3.6 𝑆2 )/(1 − 𝑆2 ) 𝑘𝑔 𝑃𝐶 + (50 − 32 𝑆2 )/(1 − 𝑆2 ) 𝑀𝐽 (7) Previous validation work has been carried out to assess the accuracy of the combustion scheme shown above which has also been incorporated into a CFD based flare model [15] applied to several gas flare systems. The first reaction represents hydrocarbon fuel combustion and describes the incomplete reaction of Fuel (F) with oxygen (O2) to produce Products of Combustion (PC) and black carbon (C) and some energy (MJ). This reaction produces S1 kilograms of black carbon per kilogram of fuel consumed where S1 depends on the fuel type (0.005 used for light hydrocarbons [42]). The second reaction represents the endothermic fuel pyrolysis or cracking reaction which produces S2 kilograms of black carbon (0.15 used for light hydrocarbons) and Intermediate Species (IS) such as carbon monoxide. The third reaction consumes black carbon and more oxygen to produce carbon dioxide (CO2) and some energy. The final reaction consumes the Intermediate Species formed in the second reaction plus some additional oxygen to form combustion products and energy. Page 18 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency This mechanism has been applied to the combustion of several flare gases as reported by Smith et.al. [42] and shown below. 2.2.1. Propane Combustion model For propane flames, a two-step chemical reaction is used that burns propane according to the following formula: C3H8 + 3.6 O2 → 3 CO2 + 1.6 H2O + 0.024 Soot + 46 MJ/kg propane (8) Soot + 2.66 O2 → 3.66 CO2 + 32 MJ/kg Soot (9) The coefficients in the formula are mass weights, not moles. The Arrhenius kinetics equation and parameters for these reactions were: Rate (moles/m3/sec) = XC2H8 * XO2 *A Exp(-Ta/T) (10) Where X is the molar concentration of the species (moles/m3), A is the pre-exponential factor (3.29x1010 for the 1st reaction and 8.0x1011 for the 2nd reaction), Ta is the activation temperature in Kelvins (15,922 and 26,500), and T is the local gas temperature (K). The characteristic time from the kinetics equation was combined with the characteristic turbulence time scale: tturb=C x2 / diff (11) Where x is the characteristic cell size, C is a user input constant (0.2x10-4), diff is the eddy diffusivity from the turbulence model, and tturb is the turbulence time scale, i.e. characteristic time required to mix the contents of a computational cell. The reaction rates are combined by simple addition of the time scales. 2.2.2. Ethylene Combustion Model For unsaturated hydrocarbon combustion, the requirement of using a pilot flame was eliminated by implementing a three-step chemical reaction. Using this approach, the ethylene combustion model consisted of the following three step mechanism: C2H4 + 0.57 O2 → 0.93 C2H2 + 0.64 H2O + 9.4 MJ/kg ethylene (12) C2H2 + 2.58 O2 → 2.7 CO2 + 0.7 H2O + 0.2 Soot + 34.1 MJ/kg intermediate (13) Soot + 2.66 O2 → 3.66 CO2 + 32MJ/kg Soot (14) As before the coefficients are mass weights, not mole weights. The Arrhenius kinetics equation and parameters for these reactions were: 1st Reaction Rate (moles/m3/sec) = XC2H4 * XO2 *1.0e15 Exp(-10,500/T) (15) 2nd Reaction Rate (moles/m3/sec) = XC2H2 * XO2 *1.0e11 Exp(-15,500/T) (16) 3rd Reaction Rate (kg/m3/sec) = YC * YO2 *1.0e11 Exp(-20,500/T) (17) Page 19 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Where X is a mole concentration (mole density) and Y is a mass concentration (partial mass density). The advantage of the three-step reaction is that the first reaction has a low activation energy, which allows the partial burning and heat release of ethylene. This will maintain combustion since the partial heat released will allow the second reaction, which produces most of the heat and all of the soot, to occur. As in the propane combustion model the ethylene Arrhenius combustion time scale is combined with the turbulence time scale to yield an overall time scale for the reaction rate. 2.2.3. Mixed Gas Combustion Model A three-step chemical reaction formulation was implemented for a mixed gas having an approximate composition of 32% C2H4, 20% C2H6, and 34% H2 (mole percent). Any remaining gases are ignored in the combustion model. The simplified 3-step reactions are: 0.572 C2H4 + 0.383 C2H6 + 0.043 H2 +0.982 O2 → 0.53 C2H2 + 0.34 C2H3 + 1.1 H2O + 14.2 MJ/kg (18) 0.61 C2H2 + 0.39 C2H3 + 2.66 O2 → 2.66 CO2 + 0.813 H2O + 0.181 Soot + 34.4 MJ/kg Soot + 2.66 O2 → 3.66 CO2 + 32 MJ/kg (19) (20) As before the coefficients are mass weights, not mole weights. The Arrhenius kinetics equation and parameters for these reactions were: 1st Reaction Rate (moles/m3/sec) = Xfuel * XO2 *1e15 Exp(-10500/T) (21) 2nd Reaction Rate (moles/m3/sec) = Xmix * XO2 *1e12 Exp(-15500/T) (22) 3rd Reaction Rate (kg/m3/sec) = YC * YO2 *1e11 Exp(-20500/T) (23) 2.3. CFD Modeling Three computations for single tip, three tips, and multi-tip flares were conducted using C3D software. The turbulent reaction chemistry coupled with radiative transport between buoyancy driven fires and surrounding objects was simulated using this tool. 2.3.1. Physical Parameters For the purposes of modeling of this study, an average gas density of 1.04 kg/m3 was assumed. Also, a standard gas pressure and temperature of 1 atm and 293.15 K was used. Moreover, an average molecular weight of 25 kg/ kmol for the volume flow for each flare was considered in the available operating data. 2.3.2. Single Tip Flare Modeling In order to obtain the flame dynamics, including flame shape and size, of a single flare tip, C3d simulation cases have been performed. Also, thermal radiation and soot estimation were found using these computations. The simulations was carried out using a 3-D physical domain with dimensions of 6 m, 6 m, and 26 m for the length, width, and height respectively. The flare tip was located 2 m above the ground level. Rectangular cells were used to construct the mesh of the Page 20 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency physical domain where the number of cells in this domain was 110,000 cells as shown in Figure 15. The 3-D domain of 30, 35, 25 m dimensions was used for simulation with radiation meters (solid boxes) a distance of 15 m and 50 m respectively from the flare burners for the three tips flare as shown in Figure 16. The mesh was refined locally near burner tips and radiation measuring unit. The total number of control volumes was 188,000 of computational cells as shown in Figure 16. Figure 15 - Single Tip flare mesh [14] Page 21 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 16 - Geometry and Mesh for 3-tips Flare [14] 2.3.3. Multi Tips Flare System There are many issues associated with multi tip flare system design. One of these issues is the difficulties of anticipated flare gas flow rates and flaring duration. Also, feed composition and its temperature conditions to flaring system are complex to specify. Moreover, flame height with respect to the fence height should be considered in the design task. Furthermore, noise and radiation to the surrounding are important in the design process. In multi-tip flares, a large plume is created from merging of all plumes of flare tips as shown in Figure 17. Figure 17 - Multi tip flare plume formation [43] Due to the large field of this kind of flare system, a large flare plume is formed, and the risk assessment associated with this large plume, high flame temperature and heat radiation is difficult Page 22 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency to perform. The flare performance quantification is also difficult to perform. Therefore, simulations with a suitable simulation package are used to quantify flare performance and impacts to surroundings (e.g., radiation to personnel and equipment, emissions, etc.) with different scenarios. The multi-tip flare field includes a wind fence that surrounds the burner tips. The main objectives of this fence are to protect workers and equipment from thermal radiation and to protect the near-burner region form high wind speed that could adversely affect the flame shape or even cause flame extinguishment. An understanding of flare tip performance and wind effects on the flame are required in order to estimate emissions from a mulit-point ground flare (MPGF). In addition, other factors important to MPGF performance can be investigated with an appropriately validated CFD model such as: the effects of tip geometry, near tip transients, burner tip interaction for hundreds of tips and plume merging and dispersion. The MPGF domain size was 35, 35, and 25 meters for the length, width, and height respectively. Figure 18 presents the mesh of the MPGF simulation which included 1.2 million cells with local refinement near burner rows and tips. To reduce computational time cost, hexahedral cells were used in all simulations. Two speeds for wind were considered, zero and 7 mph for all simulations. Flare vent gases for this study included propane and ethylene. Figure 18 - Multi-tip flare field mesh [14]. As discussed above, the C3d CFD model was validated against single and three flare tip experimental data from multiple tests. Page 23 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency 2.4. CFD fundamental models To predict emissions produced by multi-tip ground flares during transient ignition, the C3d CFD tool was used. This tool is based on a Large Eddy Simulation (LES) formulation to approximate the turbulent reacting flow system. The governing equations for this LES based CFD tool, assuming incompressible fluid flow, are given below [44]: The steady-state continuity equation is: ∂(ρui )/ ∂xj = 0 (23) where ρ is the density of the gas (mixture) and u is the three dimensional velocity vector. The momentum equation is: ∂(ρui ui )⁄∂xj = ∂P⁄∂xi + ∂τij ⁄∂xj + ρfi (24) with fi as the body forces, P as the pressure, and τij represented as the stress defined in equation 8: τij = μ(∂ui ⁄∂xj + ∂uj ⁄∂xi ) + (μB − 2⁄3 μ) ∂uk ⁄∂xk δij (25) The other governing equation that is solved is the energy equation which in C3d, this equation is: ∂ ln ρ ρ cp ∂(T)⁄∂xj = −(∇. q) − (∂ ln T) Dp Dt − (τ ∶ ∇ v) (26) where Cp is the specific heat. To resolve sub-filter scales for LES turbulence model, the Gaussian filter is used as shown in equation: G(x − r) = 〖(6/(π∆^2 ))〗^(1/2) exp(−(6(x − r)^2)/∆^2 ) (27) The following equations are used to model the kinetic energy dissipation on subgrid scales to molecular diffusion: τrij − 1⁄3 τkk δij = −2νt S̅ij (28) ̅ ∂u ̅ ∂u S̅ij = 1⁄2 ( ∂xi + ∂x j ) (29) j i with τrij as the stress tensor, S̅ij as the rate-of-strain tensor, and νt as the turbulent eddy viscosity. The eddy viscosity is approximated as the characteristic length scale times the velocity scale in most subgrid scale models as illustrated by the Smagorinsky-Lilly model: νt = (Cs ∆g )2 √2S̅ij S̅ij = ((Cs ∆g )2 |S|, Cs = Constant, ∆g = grid size Page 24 of 34 (30) Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency The equilibrium assumption was applied between energy production and dissipation of small scales in this model. The multi species conversation equations form is shown in equation (14) ∂ρmi ∂t + ∇. ρVmi = −∇. ⃗Ji + R i + Si (31) Where, mi is the mass fraction of species i, ⃗Ji diffusion flux of species i, Ri is the mass creation or depletion by chemical reactions, and Si source of mass. 3. Results and Discussion 3.1. Soot Emission During the ignition process, flare gas exits the flare tip into the atmosphere and has insufficient momentum and time to completely mix with sufficient oxygen and to fully react which results in excessive black carbon (soot or smoke) during the ignition step. Nearly all flares exhibit a characteristic puff of black smoke (unreacted black carbon) formed by incomplete combustion during the transient ignition process. Practical experience with gas flares suggest that the transient ignition process lasts approximately 10-30 seconds (from flow initiation to steady combustion). A conservative estimate for combustion efficiency during the ignition process may be approximately 50%. Given the number of flare ignition events per year and using an average gas flow and an average molecular weight derived from flare operating data, the minimum estimated emissions would be 700,000 kilograms of unreacted hydrocarbon emissions from these flares. 3.2. Single tip and three tips flares Figure 19 and Figure 21 compare experimental data of thermal radiation with those predicted by C3d considering wind speeds of 12 and 6 mph respectively. Figure 21 shows very good agreement between experimental and predicted thermal radiation from a single flare tip operated under a 6 mph average wind speed. The predicted data for a 12 mph wind appears to be underestimated when compared to experimental data (see Figure 19). This is likely because the wind speed effect is not constant and hence the amount of black carbon and associated radiated heat changes non-linearly for higher speeds. For the same position with different velocities than those shown in Figure 19 and Figure 21, the thermal radiation decreases with increasing wind speed. This may result from a cooling effect that higher wind speeds have on the flame. Also, as shown in Figure 11, as wind speed increases the flame becomes shorter and broader, due likely to increased mixing, and the shorter flames will result in less heat released than from the taller flames associated with low-wind conditions. Page 25 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 19 - Predicted versus measured thermal radiation for single tip flare (12 mph wind speed) The measured and predicted heat radiation data for flare tip sizes 0.0762 and 0.1016 meter (3 and 4 inches) is shown in Figure 21 with different tip pressures considered. The results indicate an increase in the thermal radiation with increasing burner pressure. This is expected since as the tip pressure increases, the amount of flared gas increases which will increase the thermal heat radiation. Figure 21 shows good agreement between experimental and predicted levels for thermal radiation. Figure 20 - Predicted versus measured thermal radiation for single tip flare (6 mph wind speed) The effect of wind speed on the heat radiation at a burner pressure of 2.8 psi and at a distance of 50 foot from the single burner tip is shown in Figure 22. This figure indicates a decrease in the thermal radiation with increasing the wind velocity which is related to a reduced flame temperature and/or less radiating black carbon in the flame under high wind conditions. This effect is summarized in Figure 22. Page 26 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 21 - Predicted versus measured thermal radiation for single 3" flare tip Figure 22 - Predicted impact of wind on thermal radiation flux from flare flame [15] 3.3. Multi-Point Ground Flare system The predicted flame height and heat radiation for and MPGF of 405 burner tips is presented in Figure 23. The combustion products are also shown in the same figure as transparent iso-surfaces. The heat radiation rate was found to be 61,000 and 35,000 W/m2 on the left and right walls respectively for the peak flow of flue gases. Moreover, values of 6600 and 6600 W/m2 on the left and right walls respectively, when the flow is sustainable, are obtained for the heat radiation. These cases were performed with no wind effect. Page 27 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency Figure 23 - Flame size, shape and combustion products prediction for MPGF system [14] 4. Conclusion Testing has been conducted for a single flare tip and a 3-flare tip configurations to measure the flame height and radiation flux when burning a variety of vent gases. These tests were conducted in a no-wind ambient condition as well with wind speeds ranging up to 12 mph. Wind effects on flame height showed a decrease in flame height with increasing wind speed illustrated by results from a single flare tip test where zero wind resulted in a was measured to flame height of be 15 - 16 m while the flame height in windy conditions was measured to be approximately 10.5 m high. In the windy condition, the flame was also wider and had less radiation flux than the flame in nonwind conditions. The multi-tip tests were conducted to assess flame interaction and cross lighting of adjacent flare tip burners using a single pilot. The three tip tests showed s flame height essentially the same as the single flare tip test. This indicated the flames operated independently and did not merge into a single larger/taller flame. Testing showed that as the tip pressure increased from 2.8 psi to 11.4 psi the flame height increased as did the radiant flux (3344 W/m2 to 6192 W/m2). This same behavior was observed for the larger 4" flare tip. Given the test results, a detailed CFD model was developed and used to simulate the flare flame shape and height for various flare gas flow rates, tip pressure and size and ambient wind conditions. Predicted soot levels and radiant heat flux from the single and multiple tip flare tests were compared to the measured values to validate the model. Using the validated model, predictions for single, three, and multi tip flare systems were performed using an LES based CFD model. The validated CFD model was also used to simulate a large industrial multi-point ground flare system burning approximately 260 kg/s of flare gas. Using the validated flare model, the predicted radiation from a single and three flare tip showed good agreement with the measured data. In addition, simulation of a 400 multi-tip ground flare system provided a reasonable estimate of the flame shape and flame height with the associated heat radiation profile on the surrounding wind fence and nearby equipment. Furthermore, at the maximum possible flare gas flow and at the sustained flare gas flow, the heat radiation predicted on the wind fence walls were estimated. Using the combustion simulations for this multi-point ground flare, assuming 50% combustion Page 28 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency efficiency during a 10-30 second ignition period, we estimated approximately 700,000 kg/yr of unburned hydrocarbons may be emitted from an industrial scale multi-point ground flare. ACKNOWLEDGMENTS We would like to acknowledge the Wayne and Gayler Laufer Endowment for its generous support of this work. We would also like to acknowledge Elevated Analytics and especially George Singer for their financial support of developing new flare technology that can help improve our world and the environment we all live and work in. Finally, we would like to acknowledge Zeeco for providing the opportunity to analyze several multi-point ground flares over the years which have led to much of the information presented in this paper. REFERENCES [1] The Worldbank, "Global Gas Flaring Reduction Partnership (GGFR)," [Online]. Available: http://www.worldbank.org/en/programs/gasflaringreduction#7. [Accessed 26 June 2018]. [2] C. Elvidge, M. Zhizhin, K. Baugh, F. Hsu and T. Ghosh, "Methods for Global Survey of Natural Gas Flaring from Visible Infrared Imaging Radiometer Suite Data," Energies, vol. 9, no. 14, 2016. [3] C. Elvidge, D. Ziskin, K. Baugh, B. Tuttle, T. Ghosh, D. Pack, E. Erwin and M. Zhizhin, "A Fifteen Year Record of Global Natural Gas Flaring Derived from Satellite Data," Energies, vol. 2, no. 3, pp. 595-622, 2009. [4] U.S.EPA, "EPA Enforcement Targets Flaring Efficiency Violoations; Available online.," August 2012. [Online]. Available: https://www.epa.gov/enforcement/epa-enforcementalert-epa-enforcement-targets-flaring-efficiency-violations-volume-10. [Accessed August 2012]. [5] API Standard 521, Guide for Pressure-Relieving and Depressuring Systems, 5 ed., American Petroleum Institute, 2012. [6] U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, Dec 2009). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.. [7] B. M. Conrad and M. R. Johnson, "Field Measurements of Black Carbon Yields from Gas Flaring," Environmental Science & Technology, vol. 51, p. 1893−1900, 2017. Page 29 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency [8] M. R. Johnson, R. W. Devillers and K. A. Thomson, "A Generalized Sky-LOSA Method to Quantify Soot/ Black Carbon Emission Rates in Atmospheric Plumes of Gas Flares," Aerosol Science and Technology, vol. 47, pp. 1017-1029, 2013. [9] J. D. McEwen and M. R. Johnson, "Black Carbon Particulate Matter Emission Factors for Buoyancy-Driven Associated Gas Flares," Journal of the Air & Waste Management Association, vol. 62, no. 3, p. 307-321, 2012. [10] D. T. Allen, V. M. Torres and F. C. Saldana, "Carbon Dioxide, Methane and Black Carbon Emissions from Upstream Oil and Gas Flaring in the United States," Chemical Engineering, vol. 13, pp. 119-123, 2016. [11] D. A. Lack, H. Moosmüller, G. R. McMeeking, R. K. Chakrabarty and D. Baumgardner, "Characterizing Elemental, Equivalent Black, and Refractory Black Carbon Aerosol Particles: A Review of Techniques, their Limitations and Uncertainties," Anal. Bioanal. Chem., vol. 406, pp. 99-122, 2014. [12] C. L. Weyant, P. B. Shepson, R. Subramanian, M. O. L. Cambaliza, A. Heimburger, D. McCabe, E. Baum, B. H. Stirm and T. C. Bond, "Black Carbon Emissions from Associated Natural Gas Flaring," Environmental Science & Technology, vol. 50, pp. 2075- 2081, 2016. [13] A. Wang, H. H. Lou, D. Chen, A. Yu, W. Dang, X. Li, C. Martin, V. Damodara and A. Patki, "Combustion Mechanism Development and CFD Simulation for the Prediction of Soot Emission during Flaring," Front. Chem. Sci. Eng., vol. 10, no. 4, pp. 459-471, 2016. [14] J. D. Smith, R. Jackson, A. Suo-Antilla, S. Smith and D. Allen, "Achieving Environmental Compliance through Proper Destruction Efficiency of Low-Profile Multi-Tip Flare Systems," in American Flame Research Committees-Industrial Combustion Symposium, Hyatt Regency Hotel Houston, Texas, September 2014. [15] J. Smith, A. Suo-Antilla, S. Smith and J. Modi, "Evaluation of the Air-Demand, Flame Height, and Radiation Load on the Wind Fence of a Low-Profile Flare Using ISIS-3D," in AFRC-JFRC 2007 Joint International Combustion Symposium, Marriott Waikoloa Beach Resort, Hawaii, 21 - 24 October, 2007. [16] J. Smith, A. Suo-Antilla, N. Philpott and S. Smith, "Prediction and Measurement of MultiTip Flare Ignition," in American Flame Research Committees - International Pacific Rim Combustion Symposium, Advances in Combustion Technology: Improving the Environment and Energy Efficiency, Sheraton Maui, Hawaii, September 26 -29 (2010). Page 30 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency [17] J. Smith, R. Jackson, A. Suo-Anttila, K. Hefley, D. Wade, D. Allen and S. Smith, "Prediction and Measurement of Multi-Tip Flare Ignition," in " American Flame Research Committees 2013 - Industrial Combustion Symposium, Safe and Responsible Development in the 21st Century , Sheraton Kauai, Hawaii, September 22 -25, 2013. [18] J. Smith, R. Jackson, A. Suo-Anttila, K. Hefley, Z. Smith, D. Wade, D. Allen and S. Smith, "Radiation Effects on Surrounding Structures from Multi-Point Ground Flares," in AFRC 2015 Industrial Combustion Symposium, Historic Fort Douglas Officers Club University of Utah, Salt Lake City, Utah, September 9-11, 2015. [19] D. Miller, "New Model for Predicting Thermal Radiation from Flares and High Pressure Jet Fires for Hydrogen and Syngas," Process Safety Progress, vol. 35, no. 3, pp. 237-251, 2017. [20] L. Smoot, J. Smith and R. Jackson, "Technical Foundations of New Performance Criteria for Efficient Operation of Industrial Steam-Assisted Gas Flares," in IFRF 16th International Members' Conference: Combustion and Sustainability - New Technologies, New Fuel, New Challenges, Boston, Massachusetts, June 8-10, 2009. [21] C. E. Baukal, The John Zink Combustion Handbook, Boca Raton, Florida: CRC Press, 2001. [22] D. Castineira and T. Edgar, "CFD for Simulation of Steam-Assisted and Air-Assisted Flare Combustion Systems," Energy & Fuels, vol. 20, no. 3, pp. 1044-1056, 2006. [23] J. H. Pohl, R. Payne and J. Lee, "Evaluation of the Efficiency of Industrial Flares: Test Results - EPA-600/2-84-095," US Enviromental Protection Agency, Research Triangle Park, NC, May 1984. [24] J. H. Pohl and N. R. Soelberg, "Evaluation of the efficiency of industrial flares: Flare Head Design and Gas Composition - EPA-600/2-85-106," US Environmental Protection Agency, Research Triangle Park, NC, September 1985. [25] J. H. Pohl and N. R. Soelberg, "Project Summary Evaluation of the Efficiency of Industrial Flares : H2S Gas Mixtures and Pilot Assisted Flares," EPA-600/2-86-080, September, 1986. [26] M. McDaniel, "Flare Efficiency Study - EPA-600/2-83-052," United States Enviromental Protection Agency, Research Triangle Park, NC, January 1984. [27] D. T. Allen and V. M. Torres, "TCEQ 2010 Flare Study Final Report," Texas Commission on Environmental Quality, PGA No. 582-8-862-45-FY09-04, Austin, TX, August 1, 2011. Page 31 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency [28] D. M. Leahey, K. Preston and M. Strosher, "Theoretical and Observational Assessments of Flare Efficiencies," Journal of the Air & Waste Management Association, vol. 51, pp. 16101616, 2001. [29] E. Bourguignon, M. R. Johnson and L. W. Kostiuk, "The Use of a Closed-Loop Wind Tunnel for Measuring the Combustion Efficiency of Flames in a Cross Flow," Combustion and Flame, no. 119, pp. 319-334, 1999. [30] M. R. Johnson and L. W. Kostiuk, "Effects of a Fuel Diluent on the Efficiencies of Jet Diffusion Flames in a Crosswind," in Presented at The Combustion Institute, Canadian Section, Spring Technical Meeting, Edmonton, Alberta, May 1999. [31] J. Seebold, "Practical Implications of Prior Research on Today's Questions about Flare Emissions," in American Flame Research Committee Conference-Applicatons of Combustion Fundamentals to Practical Problems, Salt Lake City, Utah, March 18-19, 2004. [32] D. Castiñeira and T. F. Edgar, "Computational Fluid Dynamics for Simulation of WindTunnel Experiments on Flare Combustion Systems," Energy & Fuels, vol. 22, no. 3, p. 1698- 1706, 2008. [33] Standards, U.S. EPA Office of Air Quality Planning and, "Parameters for Properly Designed and Operated Flares," April 2012. [34] P. E. Gogolek and A. Hayden, "Efficiency of Flare Flames in Turbulent Crosswind," in Advanced Combustion Technologies, Natural Resources Canada, American Flame Research Committee Spring Meeting, May 2002. [35] J. Pohl, P. Gogolek, R. Schwartz and J. G. Seebold, "The Effect of Waste Gas Flow & Composition Steam Assist & Waste Gas Mass Ratio Wind & Waste Gas Momentum Flux Ratio Wind Turbulence Structure on the Combustion Efficiency of Flare Flames," July 2011. [36] S. Evans, "A Critique of the Flare Provisions of EPA's proposed Refinery Sector Rule," in AFRC 2014 Industrial Combustion Symposium, Hyatt Regency Hotel, Houston, Texas, September 7-10, 2014. [37] J. Seebold, "Historial Perspective on Flare Emissions," in Texas Natural Resouce Conservation Commission Symposium, Austin, Texas, June 18, 2002. Page 32 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency [38] S. Evans, "Adventures in Flaring, what the EPA Didn't Tell You," in AFRC 2015 Indusrial Combustion Symposium, Historic Fort Douglas Officers Club, University of Utah, Salt Lake City, Utah, September 9-11, 2015. [39] 5. E. API RECOMMENDED PRACTICE 521, "Guide for Pressure-Relieving and Depressuring Systems," American Petroleum Institute, Manufacturing, Distribution and Marketing Department, Washington D.C., September 2003. [40] M. P. Company, "Performance Test of a Steam-Assisted Elevated Flare With Passive FTIR," Clean Air Engineering, Inc., Texas Refining Division, 502 10th Street South, Texas City, Texas 77590, May 2010. [41] S. Smith, J. Modi, J. M. McDonald, C. Little, J. D. Smith and L. Berg, "Environmental Testing of an Advanced Flare Tip for a Low-Profile Flare Burning Ethylene," in AFRCJFRC 2007 Joint International Combustion Symposium, Marriott Waikoloa Beach Resort, Hawaii, Hawaii, October 21-24, 2007. [42] J. Smith, A. A. Suo-Ahttila, S. Smith and J. Modi, "Estimation of the Air-Demand, Flame Height, and Radiation Load from Low-Profile Flare using ISIS-3D," in American - Japanese Flame Research Committees International Symposium, Marriott Waikoloa, Hawaii , October 22 -24, 2007. [43] J. D. Smith, R. Jackson, V. Sreedharan, A. Suo-Anti and S. Smith, "Withstanding the Wind," Hydrocarbon Engineering, vol. 21, no. 10, pp. 43-49, October 2016. [44] J. Smith, R. Jackson, V. Sreedharan, A. Suo-Antilla, Z. Smith, D. Allen, D. DeShazer and S. S., "Safe-Operation of Adjacent Multi-Point Ground Flares: Predicted and Measured Flame Radiation in Cross Flow Wind Conditions," in AFRC 2016: Industrial Combustion Symposium, Sheraton Kauai, Kauai, Hawaii, September 9-11, 2016. [45] F. S.P. and A. Hamins, "An estimate of the correction applied to radiant flame measurements due to attenuation by atmospheric CO2 and H2O," Fire Safety Journal, vol. 37, pp. 181-190, 2002. [46] J. Smith, R. Jackson, V. Sreedharan, A. Suo-Anttila, D. Allen and S. Smith, "Withstanding the Wind," Hydrocarbon Engineering, vol. 21, no. 10, p. 43 - 49, 2016. [47] R. Said, A. Garo and R. Borghi, "Soot Formation Modeling for Turbulent Flames," Combustion and Flame, vol. 108, pp. 71-86, 1997. Page 33 of 34 Prediction and Measurement of Multi-Tip Flare Ignition Advances in Combustion Technology: Improving the Environment and Energy Efficiency [48] J. D. Smith, A. Suo-Anti, R. Jackson and S. Smith, "Prediction of Plume Formation and Dispersion from Gas Flares," in 2012 Annual American Flame Research Committee Meeting, Salt Lake City, Utah, September 5 - 7, 2012. [49] J. Smith, B. Adams, R. Jackson and A. Suo-Anttila, "Use of RANS vs LES Modelling for Industrial Gas-fired Combustion," Industrial Combustion, 2017. [50] J. D. Smith, R. Jackson, A. Suo-Antilla, S. Smith and D. Allen, "Achieving Environmental Compliance through Proper Destruction Efficiency of Low-Profile Multi-Tip Flare Systems," in American Flame Research Committees- Industrial Combustion Symposium, Hyatt Regency Hotel Houston, Texas, September 2014. [51] R. Said, A. Garo and R. Borghi, "Soot Formation Modeling for Turbulent Flames," Combustion and Flame, vol. 108, pp. 71-86, 1997. Page 34 of 34