Basic Ground Flare Noise Propagation

Paper from the AFRC 2013 conference titled Basic Ground Flare Noise Propagation by Bryan Beck.

Basic Ground Flare Noise Propagation BRYAN L. BECK, P.E. Senior Flare Applications Engineer UOP Callidus 1 Contents Introduction ........................................................................................................................................... 3 Sound Basics........................................................................................................................................... 3 Figure 1: Hearing Loss vs. Age, Frequency, and Sex ............................................................................. 3 Figure 2: Sound Pressure Level (SPL) and Sound Pressure (Pa)1 ........................................................... 4 Figure 3: Equal Loudness Curves (Fletcher-Munson Curves) ................................................................ 5 Figure 4: A-Weighted Adjustments to Noise Measurement ................................................................. 6 Table 1: Permissible Noise Exposures .................................................................................................. 7 Noise Propagation .................................................................................................................................. 8 Directivity ............................................................................................................................................... 9 Figure 5: Directivity Index at Stack Exit, Large Diameter Stack8 .......................................................... 10 Atmospheric Absorption ........................................................................................................................ 10 Figure 6: Atmospheric Absorption vs. Frequency5 ............................................................................. 11 Figure 7: Vegetation Attenuation vs. Frequency7............................................................................... 12 Reflection ............................................................................................................................................. 12 Figure 8: Direct Path vs. Reflected Sound .......................................................................................... 13 Refraction ............................................................................................................................................ 13 Figure 9: Refraction of an Elevated Source ........................................................................................ 14 Barriers ................................................................................................................................................ 14 Figure 10: Ground Flare Sound Path with Solid Barrier ...................................................................... 15 Figure 11: Ground Flare Sound Path with High Air Flow Barrier ......................................................... 16 Figure 12: Transmission Loss Through Steel Plate9 ............................................................................ 16 Totally Enclosed Ground Flare (TEGF) Combustion Noise Testing ........................................................... 17 Figure 13: Totally Enclosed Ground Flare (TEGF) ............................................................................... 17 Figure 14: Digital Combustion Model, Typical Side-Fired Totally Enclosed Ground Flare (TEGF) ......... 18 Table 2: Totally Enclosed Ground Flare Measurement Points ............................................................ 19 Figure 15: Elevation View, Sound Measurement Points ..................................................................... 19 Figure 16: Plan View, Sound Measurement Points ............................................................................ 19 Table 2: Totally Enclosed Ground Flare Measurement Data .............................................................. 21 Figure 17: Totally Enclosed Ground Flare Noise Test, Steam Only ...................................................... 22 Figure 18: Comparative Noise Profile, A-Weighted ............................................................................ 23 2 Figure 19: Comparative Noise Profile, Unweighted ........................................................................... 24 Conclusions and Summary ..................................................................................................................... 25 References ........................................................................................................................................... 26 3 Introduction In recent years, global demand for ground-based flare systems has increased greatly. Ground flares have many advantages over a typical elevated flare, including reduced flame visibility and greater smokeless capacity. These two features of ground flares are well-suited to usage in populated or heavily industrialized areas where the visible flame and smoke production of an elevated flare would have a significant impact on the surrounding community. The noise principles presented in this paper are not new. This paper is an attempt to present the basic principles of environmental noise in such a way that the flare operator can make an informed decision regarding the flare technology selected. Sound Basics Sound is the result of a vibrating source generating a series of pressure waves in the air. Noise is sound from any source that results in annoyance and/or hearing loss.1 The human ear is extremely sensitive to these small pressure waves. The normal ear of a young person can detect pressures as low as 20 􀁐􅁐Pa. This represents a ratio of 1:2 x 1010 to normal atmospheric pressure.2 In other words, a pressure change of 0.000 000 02% is audible to the human ear at certain frequencies. The normal ear of a young person can hear all frequencies ranging from 20 Hz to 20,000 Hz.3 As a person ages, the ability to hear higher frequencies gradually diminishes. Hearing loss at high frequencies is more prevalent in men than women. Figure 1: Hearing Loss vs. Age, Frequency, and Sex [Harry F. Olson. Modern Sound Reproduction. Krieger Publishing Company, 1978] Because the range of human hearing encompasses such a broad range, sound is normally measured on a log-based scale known as the Decibel (dB) scale. 4 decibel = 10 Log( 􀜽􃴇􀜿􃼇􀝋􄬇􀝑􅄇􀝏􄼇􀝐􅀇􀝅􄔇􀜿􃼀􀀃􀌇􀝁􄄇􀝊􄨇􀝁􄄇􀝎􄸇􀝃􄌇􀝕􅔇 􀝎􄸇􀝁􄄇􀝂􄈇􀝁􄄇􀝎􄸇􀝁􄄇􀝊􄨇􀜿􃼇􀝁􄄀􀀃􀌇􀝁􄄇􀝊􄨇􀝁􄄇􀝎􄸇􀝃􄌇􀝕􅔩 ) Total acoustic energy is very difficult to measure; however, acoustic energy is proportional to the square of the sound pressure. decibel = 10 Log( 􀝏􄼇􀝋􄬇􀝑􅄇􀝊􄨇􀝀􄀀􀀃􀌇􀝌􄰇􀝎􄸇􀝁􄄇􀝏􄼇􀝏􄼇􀝑􅄇􀝎􄸇􀝁􄄋􀬶􃘇 􀝎􄸇􀝁􄄇􀝂􄈇􀝁􄄇􀝎􄸇􀝁􄄇􀝊􄨇􀜿􃼇􀝁􄄀􀀃􀌇􀝏􄼇􀝋􄬇􀝑􅄇􀝊􄨇􀝀􄀀􀀃􀌇􀝌􄰇􀝎􄸇􀝁􄄇􀝏􄼇􀝏􄼇􀝑􅄇􀝎􄸇􀝁􄄋􀬶􃘩) This can be restated as: decibel = 20 Log( 􀝏􄼇􀝋􄬇􀝑􅄇􀝊􄨇􀝀􄀀􀀃􀌇􀝌􄰇􀝎􄸇􀝁􄄇􀝏􄼇􀝏􄼇􀝑􅄇􀝎􄸇􀝁􄄇 􀝎􄸇􀝁􄄇􀝂􄈇􀝁􄄇􀝎􄸇􀝁􄄇􀝊􄨇􀜿􃼇􀝁􄄀􀀃􀌇􀝏􄼇􀝋􄬇􀝑􅄇􀝊􄨇􀝀􄀀􀀃􀌇􀝌􄰇􀝎􄸇􀝁􄄇􀝏􄼇􀝏􄼇􀝑􅄇􀝎􄸇􀝁􄄩 ) The reference sound pressure (zero on the decibel scale) is set at 20 􀁐􅁐Pa. Figure 2: Sound Pressure Level (SPL) and Sound Pressure (Pa)1 As the decibel scale is a logarithmic scale, the rules for adding logarithms must be followed when adding the sound pressure level (SPL) of two sources. For example, if you have two 85 dB sources side-by-side, the total SPL is 88 dBa. The formula for adding multiple sources in decibel units is: 5 􀷍􌴇􀜵􃔇􀜲􃈇􀜮􂸽 = 10􀀃􀌇􀜮􂸇􀜱􃄇􀜩􂤨(􀷎􌸱10( 􀯌􌰋􀯉􌤋􀯅􌔌􀳔􍐋 􀬵􃔋􀬴􃐩 )) 􀯡􎄋 􀯜􍰋􀭀􄀋􀬵􃕔 The human ear is not equally sensitive across the entire audible spectrum. At typical office volume, the ear is significantly more sensitive to sound in the region from 2,000 to 5,000 Hz than at other frequencies as is shown in the following graph. Figure 3: Equal Loudness Curves (Fletcher-Munson Curves) (R. Nave. "Equal Loudness Curves" hyperphisics.phy-astr.gsu.edu. Retrieved July 12, 2013, from http://hyperphysics.phy-astr.gsu.edu/sound/eqloud.html) In order to compensate for variations in sensitivity across the audible spectrum, a weighting factor is normally applied to SPL measurements taken in industrial environments. The most common is "A" weighting. Measurements adjusted by the "A" weighting are normally referred to in units of dBA. By using A-weighting to adjust SPL measurement, the resulting measurement more accurately reflects the response of the human ear. 6 Figure 4: A-Weighted Adjustments to Noise Measurement In order to limit hearing damage in the United States, the US Occupational Health and Safety Administration (OSHA) has issued regulations concerning noise exposure in the workplace. These regulations establish dose limits based on sound pressure level vs. exposure time. Similar regulations exist in other countries. The OSHA standard effectively limits long term personnel noise exposure to <90 dBA. 7 Table 1: Permissible Noise Exposures (29 CFR 1910.95(b)(2) Table G-16) Duration per Day, Hours Sound Level dBA Slow Response 8 90 6 92 4 95 3 97 2 100 1 ½ 102 1 105 ½ 110 ¼ or Less 115 The total noise dose for an individual can be calculated using the equation: %􀜦􂘇􀝋􄬇􀝏􄼇􀝁􄄽 = 􀷎􌸇 􀜥􂔋􀯜􍰇 􀜶􃘋􀯜􍰋 􀯡􎄋 􀯜􍰋􀭀􄀋􀬵􃕆 For this equation, Ci is the actual time exposed to each dB level and Ti is the actual time exposed at each noise level. In addition to the maximum limits, a dose in excess of 50% will be placed in a hearing conservation program. For example, exposure to 90 dBA in excess of four hours would exceed 50% dose and would require a hearing conservation program for the employee. There are several employer requirements associated with a hearing conservation program, including a requirement for annual employee audiograms provided free of charge by the employer. Additional studies have shown that continuous noise exposure to 70 dBA or less over a forty-year working life would produce virtually no hearing loss. Exposure to 75 dBa or less for eight 8 hours per day over a forty-year working life was found to be acceptable. Both criteria are very restrictive and have not been incorporated into OSHA regulations.4 Flare specifications are often written that require flares to operate at 85 dBa or less at the maximum rate. For an elevated flare designed for emergency service, designing the flare for 85 dBA or less at the maximum emergency rate might result in greatly increased flare cost for very little value. First, an emergency flare will only operate at high rates for short periods of time. Second, the radiant heat generated by an emergency flare at high waste gas flow rates will generally force personnel to evacuate the area long before noise damage occurs. However, efforts to reduce the noise profile of a flare system at sustained rates could provide more environmental benefit for the cost incurred. Noise Propagation Noise propagation in an industrial environment is affected by geometrical spreading, atmospheric refraction, attenuation and reflection due to barriers, atmospheric absorption, and attenuation from intervening vegetation.5 The noise field emitted from a flare is generally split into two fields: far field and near field. In the far field, the noise source is normally modeled as point source. The point source approximation is often applied to elevated flare tip because the elevated tip will usually be a significant distance from a measuring location. Noise propagation from a point source undergoes spherical spreading. During spherical spreading the sound pressure is proportional to the area of the sphere. Therefore, for each doubling in distance from the point source the sound pressure level will be reduced by 6 dB.6 Sound calculations in the near field will be much more difficult. In the near field, there may be multiple noise sources or a large noise source spread over some distance.6 Noise propagation is not spherical in the near field. For example, if a noise monitor were mounted on an elevated platform near a flare tip, noise would not appear to be spreading spherically. Instead, the noise would be emitted from multiple locations, including: 􀁸􇡊 Jet noise at the flare exit 􀁸􇡃 Combustion noise along the length of the flame 􀁸􇡊 Jet noise from assist steam For flare equipment near field noise is extremely difficult to calculate. For elevated flares near field noise calculations are normally not of concern. Anyone standing in the near field of an elevated flare would likely be injured by the flame long before the noise because a safety concern. 9 The near field noise is much more important for a ground flare. A ground flare has numerous noise emitters, reflectors, and absorbers during normal operation. The staged nature of a typical ground flare results in different noise emitter configurations as burners are placed into and removed from service. A location that measures 85 dBA may be just a few feet from a location that measures 3 dBA higher or lower. Ground flare specifications are often written requiring sound pressure limits at one meter from the equipment. This is within the near field zone of most large ground flares. Additionally, most ground flare radiant barriers are constructed to allow combustion air to flow through the barrier. A barrier that is designed to maximize air flow is going to have reduced impact on noise flow and can produced localized areas of elevated noise. The area around a ground flare will not normally be occupied during heavy flaring. It is recommended that ground flare noise measurements are recorded at least ten meters from the fence to minimize localized effects. If there are numerous reflective surfaces within the far field a reverberant field may result. An example of a reverberant field would be a flare noise measurement inside a tank farm. In the reverberant field the sound does not spread spherically and instead reflects back upon the measurement point. For each doubling of distance inside a reverberant area the sound pressure will drop 3 dBA due to spreading and reflection.6 Directivity Directivity is a measure of the variation of a source's sound radiation with direction. It varies with the angular position around the source and the frequency being measured. Directivity has a strong influence on the jet noise component of high velocity flare tips and steam assist systems. Directivity can result in noise reductions of 10 dBA or greater at frequencies higher than 1,000 Hz. 10 Figure 5: Directivity Index at Stack Exit, Large Diameter Stack8 Atmospheric Absorption Air is not a perfect transmitter of sound. As sound energy passes through the air it is absorbed by the gasses that make up the atmosphere. The absorption of sound is particularly pronounced at high frequencies but is almost negligible at low frequencies. 11 Figure 6: Atmospheric Absorption vs. Frequency5 Ground cover and trees also serve to attenuate noise. The noise is attenuated by reflection and absorption of the leaves. While vegetation is normally limited inside an industrial facility, vegetation is often left as a visual screen between an industrial facility and surrounding residential areas. Vegetation provides very good attenuation of high frequency noise but its impact on low frequency noise is limited. Deciduous trees provide an insignificant amount of noise reduction during the winter months. The graph below is a sample plot of vegetation attenuation versus frequency. Actual vegetation attenuation will vary widely based on the type, thickness, and height of the vegetation.7 12 Figure 7: Vegetation Attenuation vs. Frequency7 Reflection Sound waves will reflect from hard surfaces. Different surfaces reflect noise differently. Surfaces such as grass or snow tend to absorb more sound and reflect less than surfaces such as gravel or water. An observer close to the ground could receive both direct path sound waves and reflected sound waves from the ground. However, this rarely doubles the amount of sound received because the reflected sound wave is rarely perfectly in phase with the direct path sound wave. Under certain conditions, the reflected sound wave will be near 180􀁱􇄠 out-of-phase with the direct path sound wave, resulting in partial cancellation of the direct path noise near grade. In the United States and Canada the regulatory receiver height is 1.5 meters, which means the receiver could be in the cancellation zone. In Europe, the regulatory receiver height is 4 meters, which means the receiver will always be above a potential noise cancellation zone.5 13 Figure 8: Direct Path vs. Reflected Sound Refraction The speed of sound (a) in dry air at 0􀁱􇄠 C and one atmosphere is 330 meters per second. The speed of sound in air is proportional to the square root of the temperature.10 􀜽􃴅 􀗟􍼃 􀎾􋸇􀜶􃙓 Sound waves will travel in a straight line in a homogenous atmosphere. In an outdoor industrial environment the atmosphere is rarely homogenous. There will normally be wind, temperature, and humidity variations across a given vertical column of air. As sound passes through these variations, the local speed of sound will vary, resulting in curving of the sound propagation path. The following figure represents a flat location with solar heating of the ground. Solar heating results in higher temperatures near grade causing the sound to curve upwards. The lower arc is the limiting arc and just grazes the ground before curving upwards. The curvature of the arc creates a shadow zone in which flare noise is only audible through diffraction or reflection from adjacent structures.5 14 Figure 9: Refraction of an Elevated Source A problematic case is a negative temperature inversion in which the temperature increases with increasing elevation. During temperature inversion conditions, the flare sound path curves downward towards grade. Under ideal conditions, the noise reflects upward from grade and curves back towards grade. This is particularly noticeable over water and other flat, hard surfaces. The sound cannot escape upwards and only spreads over two dimensions rather than three. This results in noise propagation over extremely long distances. This effect is not limited to temperature inversions. Wind at high elevations can also refract sound downward.5 Barriers Any large and dense object that blocks the path of a sound wave is considered a barrier.5 Some examples include a concrete wall, a tank, a building, or a solid radiation barrier around a rectangular ground flare. These barriers create a shadow zone. Sound does not have a direct path into the shadow zone and must enter the zone by diffraction around the barrier, diffraction over the barrier, or transmission through the barrier. Dense barriers are typically defined has having a mass per unit area of 20 kg/m2 or greater excluding framing. A dense barrier will have transmission loss in excess of 25 dB at 500 Hz.5 For engineering purposes direct sound transmission through the dense barrier can be ignored. The amount of attenuation due to diffraction over a barrier is primarily determined by the Fresnel number. 􀜰􃀽 = 2􀝖􅘇 􀟣􎌀 15 In this equation, N = Fresnel Number, z = increase in path link due the barrier, and 􀁏􄼠 is the wavelength of the sound to be considered. Taller intervening barriers result in longer path lengths, increasing the noise attenuation. Low frequency sound has a longer wavelength than high frequency sound and is less impacted by intervening barriers. Figure 10: Ground Flare Sound Path with Solid Barrier The figure above is a simplified version of the sound propagation path over a rectangular ground flare radiation barrier. In a large rectangular ground flare, large amounts of combustion air are required for adequate destruction of the waste gas and to maintain acceptable temperatures within the unit. The radiation barrier is normally designed to allow some air to flow under and through the fence, reducing the effectiveness of the barrier for noise blockage. Most ground radiation barriers are designed to be as lightweight as possible to reduce materials costs. A typical ground flare radiation barrier would have a complex surface pattern and a mass per unit area of 15-25 kg/m2 excluding structural bracing. Although the openings in the fence reduce barrier attenuation, the barrier does reflect the sound and attenuate the sound via transmission losses, particularly at high frequencies. 16 Figure 11: Ground Flare Sound Path with High Air Flow Barrier Figure 12: Transmission Loss Through Steel Plate9 17 Totally Enclosed Ground Flare (TEGF) Combustion Noise Testing UOP Callidus designs, manufactures, and installs cylindrical enclosed-flame flares. They are marketed as Totally Enclosed Ground Flares (TEGFs). Major components of a TEGF normally include: 􀁸􇡓 Self-supporting, refractory-lined, open top combustion cylinder. 􀁸􇡓 Stainless steel high stability burner system. The burners may be side fired or bottom fired. 􀁸􇡌 Lower combustion air windows at the bottom of the cylinder. The majority of the combustion air is drawn into the unit via natural draft in the combustion cylinder. 􀁸􇡌 Lower wind fence. The lower wind fence blocks crosswind through the unit, protects the surroundings from direct radiation from the flame, and reduces medium and high-frequency jet noise. 􀁸􇡓 Staging control system. The staging control system matches the numbers of burners on service with the system demand. 􀁸􇡃 Continuous burner pilots. 􀁸􇡂 Burner smokeless assist system. At low firing rates, small amounts steam, air, or assist gas is normally injected into the flame to promote mixing and eliminate smoke. At high firing rates the assist system is not normally needed for smokeless operation. 􀁸􇡁 Access ladders, platforms, and sampling points as required. Figure 13: Totally Enclosed Ground Flare (TEGF) 18 Figure 14: Digital Combustion Model, Typical Side-Fired Totally Enclosed Ground Flare (TEGF) UOP Callidus was requested to participate in the site noise acceptance testing for a TEGF. The unit was a steam assisted unit. The early-stage burners had a steam injection system that injected steam into the flame bundle. Steam injection forces air into the flame bundle, ensuring existing sufficient oxygen for combustion is available within the flame. Without sufficient oxygen, the free carbon is emitted as smoke. The steam system was noise-tested without combustion up to 2.5 Bar. UOP Callidus coordinated with the customer and a third-party noise testing service to conduct sound testing. Noise data was taken at four different locations around the flare as is shown in the following tables and figures. 19 Table 2: Totally Enclosed Ground Flare Measurement Points Distance Elevation Measurement Point 1 43m from Centerline 50m Measurement Point 2 43m from Centerline 7m Measurement Point 3 43m from Centerline 7m Measurement Point 4 180m from Centerline 7m Figure 15: Elevation View, Sound Measurement Points Figure 16: Plan View, Sound Measurement Points All measurements were A-weighted SPL measurements. Background noise was the first measurement. Subsequent A-weighted noise measurements were performed in the steam-only condition and at multiple waste gas flow rates. 20 With a known background level, SPL was calculated for each sampled case. MP1 was found to be the loudest location despite being further from the burners than MP2 or MP3. Possible explanations include: 􀁸􇡄 Directivity effects on the flow and combustion noise at the stack exit. 􀁸􇡈 Heavy refraction of the burner jet noise and combustion noise. o Upward velocity of the combustion products would tend to refract the noise upwards o Higher temperatures between the burners would cause the burner jet noise to refract upward as it traveled laterally across the unit. 􀁸􇡌 Less noise diffusion of the noise exiting the upper portion of the stack. Noise from the lower portion of the stack would be required to propagate through a more tortuous path through the combustion windows, past the burners and piping, and over the wind fence. 21 Table 2: Totally Enclosed Ground Flare Measurement Data Waste Gas (tons/hr) Steam (tons/hr) MP1 MP2 MP3 MP4 - - 70 64 63 61 Background Measurement - 28 78.2 74.0 72.2 61 2.5 Bar Steam Pressure Steam Only - 2 78.0 73.0 72.2 61 2.0 Bar Steam Pressure Steam Only - 19.5 77.4 72.0 72.0 61 1.5 Bar Steam Pressure Steam Only 26.5 7 78.0 71.5 72.8 62 72.5 9.5 79.6 73.0 74.2 65 82 11 79.8 72.5 73.4 62.5 96 12 80.3 72.5 72.8 63.5 107 15 80.1 72.0 72.9 63 Burners Forced Unstable With Steam Injection 20 tons/hr Waste Gas Not Recorded 85.4 - 85.5 74 Strong Vibrations and Oscillations Burners at Stability Limit 100 tons/hr Waste Gas Not Recorded 83.4 - 80.5 - Burners at Stability Limit 100 tons/hr Waste Gas Not Recorded 81.0 - 76.5 - Three steam system noise tests were performed at various flow rates. Steam system noise was much less than background noise levels below 500 Hz. Background noise was found to be inconsequential for all combustion measurements. 22 Figure 17: Totally Enclosed Ground Flare Noise Test, Steam Only The most noteworthy portion of the combustion data plot is the noise impact of flame instability. When the flame was purposely driven unstable through excess steam injection, the burners would oscillate between lit and unlit at 34 to 39 Hz, creating a heavy resonance effect in the combustion cylinder. Even though the waste gas flow rate was relatively low, the noise level was very loud in the 31.5 Hz frequency band. The high noise was easily corrected by reducing the steam flow to the burners until the burners became stable. Two additional data sets were added to the combustion data plots. The first is the API 537 combustion noise spectrum data set. The API 537 noise spectrum is based on the original work by T.J. Smith and J. K. Kilham. "Noise Generation by Open Turbulent Flames", The Journal of the Acoustical Society of America, Volume 35, Number 5. May 1963. The original noise spectrum research was based on very small flame jets 9.5mm (0.375") in diameter and smaller. It correlates well above 500 Hz, but does not appear to correlate well at lower frequencies. The second data set is a typical jet engine backblast noise spectrum.11 Far more research has been done regarding jet engine backblast than has been done on continuous flaring. The A-weighted jet noise spectrum also does not closely correlate with the ground flare noise spectrum. 23 Note that the test measurements were A-weighted. Most governmental industrial hygiene regulations are based on A-weighting; therefore, most noise specifications and equipment guarantees are based on A-weighting. Figure 18: Comparative Noise Profile, A-Weighted For comparison purposes, the combustion data was replotted with the A-weighting removed. The jet noise profile now appears more similar to the enclosed flare noise profile. It is also apparent that the combustion noise is very loud below 500 Hz, particularly when the flare is driven unstable. Low frequency noise emitted from flares will have no direct health impact on the surrounding community. Because the human ear is relatively insensitive to low frequency vibration the low frequency noise required to cause hearing damage is unrealistically high at any significant distance from the flare. However, low frequency noise can have indirect impacts on community health. The most common complaint regarding low frequency noise in the vicinity of airports is rattle of building elements and household items.11 In many cases, airports have performed extensive modifications to residential housing around the runway. Modifications included window replacement with multiple-pane windows, increased weather stripping 24 around doors and windows to stop rattle, and additional attic insulation. Cost is typically in the range of $15,000 to $30,000 per house.11 Figure 19: Comparative Noise Profile, Unweighted This raises an important point. There aren't many actions that can be done to reduce low frequency noise in residential areas around flare systems. The first solution and most obvious solution would be to flare less. The second solution is driven by the results of the steam instability test. The flare burner assist gas system must be operated correctly. If the flare is operated incorrectly the system will produce extreme low frequency noise. A third solution would be burner design; however, there are physical limitations regarding what can be accomplished via burner design. Consider a modern commercial jet engine. Millions of dollars of dedicated research and development are expended for each new jet engine design to minimize the low frequency rumble. Even with such a massive expenditure, airports are paying for attic insulation in surrounding homes because the low frequency noise still exists. Even a medium-sized rectangular ground flare (500,000 kg/hr) burns more fuel at emergency rates than dozens of fully-loaded Boeing 777's at takeoff. Many low frequency noise reduction solutions have been tried at airports: 25 􀁸􇡔 Trees and shrubs are ineffective noise attenuators at low frequencies. 􀁸􇡁 Air absorption is minimal for low frequency noise. 􀁸􇡂 Barriers close the source (such as a flare radiant barrier) are ineffective as low frequency noise simply diffuses over or through the barrier. 􀁸􇡂 Barriers can be effective if placed close to the receiver. For example, a 15 foot barrier located within 50-100' of a residence provides some low frequency attenuation within the shadow. One possible solution would be residential-scale active noise cancellation. Active noise cancellation is commonly used in stereo headsets for travelers. Active noise cancellation works by generating a sound signal that is out of phase with the sound signal from a noise source. By overlaying two waves that are out of phase, the noise signal is reduced or eliminated. This technology is commonly used in noise-cancelling headphones. Conclusions and Summary As is shown in the aforementioned data set, properly operating flare systems can produce a significant amount of low-frequency combustion noise. Current industry references might underestimate the amount of low-frequency noise produced. It is well established that low frequency noise can travel extremely long distances with much less attenuation than high frequency noise. Airport and traffic studies have demonstrated that low frequency noise will not produce hearing damage, but it can result in significant quality of life issues in the surrounding community. When placing a ground flare within a residential area, the impact of low frequency noise upon residential areas should be closely evaluated. 26 References 1. Patrick N. Breysse and Peter S. J. Lees, "Noise", Johns Hopkins School of Public Health, 2006 2. Leo L. Beranek, "Basic Acoustical Quantities: Levels and Decibels" in Istvan L. Ver and Leo L. Beranke (Ed.), Noise and Vibration Control Engineering Principles and Applications, Wiley, New Jersey, 2006 3. Sebastian Haskel and David Sygoda. Biology, A Contemporary Approach. New York: Amsco, 1996 4. Suzanne D. Smith, Charles W. Nixon, and Henning E. Von Gierke, "Damage Risk Criteria for Hearing and Human Body Vibration" in Istvan L. Ver and Leo L. Beranke (Ed.), Noise and Vibration Control Engineering Principles and Applications, Wiley, New Jersey, 2006 5. Ulrich J. Kurze and Grant S. Anderson, "Outdoor Sound Propagation" in Istvan L. Ver and Leo L. Beranke (Ed.), Noise and Vibration Control Engineering Principles and Applications, Wiley, New Jersey, 2006 6. William W. Lang, George C. Maling, Jr., Matthew A. Nobile, and Jiri Tichy, "Determination of Sound Power Levels and Directivity of Noise Sources" in Istvan L. Ver and Leo L. Beranke (Ed.), Noise and Vibration Control Engineering Principles and Applications, Wiley, New Jersey, 2006 7. C. M. Kalansuriya, A. S. Pannila, and D. U. J. Sonnadara, "Effect of roadside vegetation on the reduction of traffic noise levels", Proceeding of the Technical Sessions 25, 1-6, Institute of Physics Sri Lanka, 2009 8. Martin Hirschorn, Noise Control Reference Handbook, 1989 Edition. Industrial Acoustics Company, 1989 9. Istan L. Ver, "Interaction of Sound Waves with Solid Surfaces" in Istvan L. Ver and Leo L. Beranke (Ed.), Noise and Vibration Control Engineering Principles and Applications, Wiley, New Jersey, 2006 10. Michael R. Lindeburg, Mechanical Engineering Reference Manual, Twelfth Edition. PPI Belmont, California. 2006 11. Ben H. Sharp, Yuri A. Gurovich, and William W. Albee, "Status of Low-Frequency Aircraft Noise Research and Mitigation", Wyle Acoustics Group, Arlington, Virginia, 2001 12. N. D. Narasimhan, "Predict flare noise", Hydrocarbon Processing, April 1986 13. T. J. B. Smith and J. K. Kilham, "Noise Generation by Open Turbulent Flames", The Journal of the Acoustical Society of America, Volume 35 Number 5, May 1963 14. American National Standards Institute/American Petroleum Institute, "Flare Details for General Refinery and Petrochemical Service, " ANSI/API Standard 537 Second Edition, Washington, D.C., December 2008