Recent Developments and Current Insights in Ethylene Flare Technology: The Impact of Burner Design Vent Gas Variability and Ancillary Equipment

Paper from the AFRC 2016 conference titled Recent Developments and Current Insights in Ethylene Flare Technology: The Impact of Burner Design Vent Gas Variability and Ancillary Equipment

Pressure-assisted burners are used within multipoint flares in large numbers to provide smokeless operation without the use of steam or assist air. In the United States of America flares must conform to the regulations in 40 CFR 60.18 and 40 CFR 63.11 (40 CFR).; General observation has led to the conclusion that burner design and vent gas variability play a role in the performance of the system with regard to the thermal radiation emitted from the flame, the cross-lighting capability of the flare burners, the stability of the flame and the longevity of the flare fence and ancillary equipment.; Previous evaluation of flares for the region of stable operation has resulted in the use of exit velocity as the means of determining a stable flame. Pressure assisted flares violate the exit velocity criterion long used for flare flame stability determination, and as such the velocity limit method for determining DRE and flame stability cannot be applied.; As previously published and again reviewed for this work, neither combustion zone net heating value nor combustion zone flammability limit is a good predictor of combustion efficiency for pressure-assisted flares. While the style of the burner head is likely also significant in determining the combustion efficiency of the flare, all of the data available shows all burners of commercial quality outperformed current EPA requirements.; Analysis of thermal radiation data for a burner head of alternate design versus that of conventional design shows that for a given heat release, thermal radiation can be reduced by 15% with 95% confidence. Similarly a 49% reduction in flame length is possible with the same 95% confidence. The variability in the flame height for the burner head of alternate design is also reduced for a given vent gas flow rate.; A 16% reduction in the heating value required for cross lighting of burner flames is demonstrated through physical testing for a new burner of alternate design.; Further physical testing and data incorporated from public sources show it is plausible that the mole weight of only the flammable components of the vent gas provides an approximation to the maximum inert fraction of the vent gas allowable for stable flame.

Recent Developments and Current Insights in Ethylene Flare Technology: The Impact of Burner Design, Vent Gas Variability and Ancillary Equipment Matthew Martin, Bryan Beck Honeywell UOP Callidus Technologies LLC Tulsa, OK USA Email: [e-mail] ABSTRACT Pressure-assisted burners are used within multipoint flares in large numbers to provide smokeless operation without the use of steam or assist air. In the United States of America flares must conform to the regulations in 40 CFR 60.18 and 40 CFR 63.11 (40 CFR). General observation has led to the conclusion that burner design and vent gas variability play a role in the performance of the system with regard to the thermal radiation emitted from the flame, the cross-lighting capability of the flare burners, the stability of the flame and the longevity of the flare fence and ancillary equipment. Previous evaluation of flares for the region of stable operation has resulted in the use of exit velocity as the means of determining a stable flame. Pressure assisted flares violate the exit velocity criterion long used for flare flame stability determination, and as such the velocity limit method for determining DRE and flame stability cannot be applied. Figure 1- Burners of conventional design inside a typical multipoint ground flare. As previously published and again reviewed for this work, neither combustion zone net heating value nor combustion zone flammability limit is a good predictor of combustion efficiency for pressure-assisted flares. While the style of the burner head is likely also significant in determining the combustion efficiency of the flare, all of the data available shows all burners of commercial quality outperformed current EPA requirements. lighting of burner flames is demonstrated through physical testing for a new burner of alternate design. Analysis of thermal radiation data for a burner head of alternate design versus that of conventional design shows that for a given heat release, thermal radiation can be reduced by 15% with 95% confidence. Similarly a 49% reduction in flame length is possible with the same 95% confidence. The variability in the flame height for the burner head of alternate design is also reduced for a given vent gas flow rate. 1. INTRODUCTION A 16% reduction in the heating value required for cross Further physical testing and data incorporated from public sources show it is plausible that the mole weight of only the flammable components of the vent gas provides an approximation to the maximum inert fraction of the vent gas allowable for stable flame. Multipoint ground flares (MPGF) are depressurizing devices used for the safe disposal of vented gas streams. Figure 1 shows a typical multipoint ground flare. A high pressure vent gas stream is supplied to the burners which in turn use the pressure energy to entrain air. The resulting fuel-air mixture allows the vent gas to be flared without producing visible smoke. Alternate technologies which conform to 40 CFR exit velocity requirements utilize fans or steam to provide the required motive force and mixing energy to the combustion air stream. Some stages of a MPGF may be steam or air assisted if the vent gas pressure energy is not available. Vent gas within the flare is distributed to multiple pipes commonly referred to as ‘runners'. Each runner supplies waste gas to the burners and is controlled by an automated staging valve. As waste gas flow rate increases, individual staging valves open, © 2016 Callidus Technologies LLC, All Rights Reserved 1 placing additional runners and burners on service. This automatically matches flare capacity to system demand. By matching flare capacity to waste gas flow rate, the high pressure vent gas arrives at each burner head with the desired fuel-air mixing and smokeless flame. Each runner will typically have one or a few piloted burners to initiate the flame. Once the piloted burners are lit when the stage is brought into operation, the rest of the burners from that stage will light from those burners already in operation. The ability of the burners to cross-light and produce a stabile flame in the face of vent gas variability, particularly vent gas with inert constituents, is of critical importance. Additionally, the burners must be able to remain stable as the waste gas supply pressure increases. For any vent gas flaring technology, a certain fraction of the heat energy is transferred from the flame in the form of thermal radiation. Due to the potentially large volume of vent gas that must be disposed during a plant startup or shutdown the resulting thermal radiation emitted from the flame can cause a rapid temperature rise and damage to the surroundings if left unmitigated. Elevated flares move the center of radiation from the flame a suitable distance above grade in order reduce the resulting radiation to grade. MPGFs mitigate the radiation from the flame to the surrounding by using a thermal barrier fence, which blocks the transmission of radiant energy. In either case, the faction of heat that is released as thermal radiation is critically important to the design of the flare system. An MPGF barrier fence typically ranges from 35' to 60' in height. Additionally, the fence protecting the outside environment from the flaring within it partially shields the burner flames from wind. The pressure supplied an MPGF burner results in an exit velocity that is well above the regulated maximum velocity for a flare within the regulations of the USA. Therefore, one must file for an Alternative Means of Emissions Limitation (AMEL) in order to utilize this style of flare. To meet the requirements for a flare burner, one must demonstrate that it will cross-light, provides a stable flame over the required operating range, provides the required destruction efficiency for the application - commonly 95% or 98% -- does not produce visible smoke emissions except for the allowed five minutes during any two consecutive hours, and has flame visibility that is consistent with requirements for radiation protection and community considerations. 2. PRELIMINARIES AND REVIEW 2.1 Combustion Zone Net Heating Value and Combustion Zone Lower Flammability Limit The combustion zone net heating value (NHV cz) and combustion zone lower flammability limit (LFLcz) have been shown to correlate with combustion efficiency (CE) for low exit velocity flares[1]. In recent AMEL documents, the usage of these parameters appears to be related to flame stability with a threshold value being set to determine stable burner flame. While this may be a permissible practice if the intent is to monitor heating value of the vent gas stream and reduce the inert fraction by providing combustible assist gas, it should be realized that a strict heating value limit is not applicable for all vent gas compositions or burner types. From published data the destruction efficiency of the pressure assisted flares is greater than 99% when the flame is present but that LFLcz and NHVcz are not adequate indicators of flame stability. The following modified from "Parameters for Properly Designed and Operated Flares" is used to calculate the NHVcz [1]: Equation 2.1.1 𝑵𝑯𝑽𝒄𝒛 = 𝑸𝒗𝒈 𝑵𝑯𝑽𝒗𝒈 (𝑸𝒗𝒈 + 𝑸𝒔) Where: NHVcz (Btu/SCF) = the net heating value of the combustion zone gas Qvg (SCF/hr) = the volumetric flow rate of the vent gas NHVvg (Btu/SCF) = the vent gas net heating value Qs (SCF/hr) = the total steam rate When there is no assist steam for a vent gas, the equation reduces to the net heating value of the vent gas. For LFLcz, the equation adjusted for unit consistency from the same paper follows; the percent in the denominator is changed to mole fraction: Equation 2.1.2 𝐿𝐹𝐿𝑐𝑧 = 𝑥 ∑𝑛𝑖=1 𝑖 𝐿𝐹𝐿 1 𝑖 − (𝑁𝑒,𝐻2𝑂 − 1)(𝑥𝐻2𝑂 ) − (𝑁𝑒,𝐶𝑂2 − 1)(𝑥𝐶𝑂2 ) Where: LFLcz = lower flammability limit of combustion zone gas n= number of components in the combustion zone gas i= index of the individual component xi (mol frac.) = mole fraction of component i in the combustion zone Ne,H2O= Coefficient of nitrogen equivalency for H2O xH2O= mole fraction of water in the combustion zone Ne,H2O= Coefficient of nitrogen equivalency for H2O xCO2= mole fraction of water in the combustion zone Equation 2.1.3 𝐶𝐸% = 𝐶𝑂2 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 × 100 𝐶𝑂2 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 + 𝐶𝑂𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 + 𝐻𝑦𝑑𝑟𝑜𝑐𝑎𝑟𝑏𝑜𝑛𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 Where: CE(%) = combustion efficiency CO2concentration = concentration of CO2 CO2concentration = concentration of CO Hydrocarbonconcentration = concentration of Hydrocarbon © 2016 Callidus Technologies LLC, All Rights Reserved 2 2.2 Exit Velocity and Flame Stability For regulation compliant flares that do not require an AMEL the exit velocity of a flare is limited to 400 ft/s or less for lower heating value gas per 40 CFR 60.18 and 40 CFR 63.11 [2]. At design rate, a MPGF burner typically has choked flow and a Mach 1 exit velocity, so cannot comply with this requirement. However, much like 40 CFR compliant flares, it appears that MPGF burners are stable as long as a certain minimum calorific content of vent gas is maintained [3, 4, 5]. 2.3 Burner Operation and Flame Stability MPGF burners that flamed-out in the prior emissions testing did so under conditions wherein the NHVcz and the LFLcz were well within previously suggested limits. According to EPA document EPA-HQ-OAR-2014-0738-0002 the flame out occurred between two and 17 minutes [5]. It has been noted in private conversation and the personal observation of the author that in most cases it is plausible the test apparatus may in fact be causing variation in vent gas stream, which causes flame-out when operating close to the low calorific value stability limit of the burner. 2.6 Review of Prior Test Data for NHVcz and LFLcz Both the USA Environmental Protection Agency (EPA) docket EPA-HQ-OAR-2014-0738 and the technical memorandum of EPA-HQ-OAR-2014-0738-0002 provide flare burner performance data. This data used in combination with newly generated data was used to calculate CE sensitivity to NHVcz and LFLcz. The data table used for this analysis is given in the appendix. A general linear model ANOVA was used to analyze the variance of CE resulting in a p-value for NHVcz and LFLcz of 0.762 and 0.809 respectively. The high p-values suggest that neither value has a statistically significant effect on CE. Vent gas nitrogen content taken in combination with LFLcz yields slightly better results with an R2 statistic of 19.2 and an s statistic of 0.34. The lack of readily available input variables to explain CE may be interpreted by referring to Figure 3 where it can be seen that the CE response is essentially flat with respect to NHVcz /NHVvg-lfl: If the flame is present there is a high CE and so other explanitory variables are not required. 2.4 Burner Design, Cross-Lighting and Flame Stability The authors of Federal Register document 80 FR 8023 state that "flare head design can influence the flame stability curve" [2]. Small changes in burner design can have a large effect on the burner design's ability to cross-light and maintain a stable flame. For example conventional commercially available flare burners have specific aerodynamic features that enhance the burner stability, which may include the provision for regions of fuel-air recirculation, mixing and stagnation. Cross-lighting can be achieved through the appropriate combination of specially disposed vent gas ports on the burner heads and spacing between the burners. 2.5 Burner Design and Smokeless Operation Even with equivalent vent gas composition and pressure, smokeless operation is not ensured. The burner design has a significant impact on the pressure at which smokeless operation occurs for a given burner design. Figure 2 compares two flames emanating from burners that both have a radial arms and a central hub. The left-hand panel shows the resultant flame and smoke from a 1991 vintage burner while the right-hand panel shows the flame and smoke from a 2014 vintage burner, showing that two burners of similar description can perform in a substantially different manner. Figure 3 - Pressure-assisted flare data plotted against data from [1]. The blue diamonds represent stable steam or air assisted flares. The red boxes represent unstable steam or air assisted flares. The orange circles represent stable nonCallidus pressure assisted burners. The yellow circles represent unstable non-Callidus pressure assisted burners. The light green triangles represent stable Callidus MP4U burners. 2.7 Review of Prior Data for Burner Type Figure 2 - Comparison of the differing flame quality and smoke production from two burners of essentially the same description. The publicly available pressure assisted flare data inputs are not held constant across manufacturers. The Honeywell UOP Callidus MP4U burner test data available, which is arguably comparable, only utilizes high heating value gas with no inert content. The data for the other burner has a large range in vent gas heating values. A statistically significant difference in burner types is suggested from a two-sample t-test utilizing only the high heating value data, which resulted in a p-value of 0.022. The box plot of Figure 5 also shows the MP4U to be more consistent in performance with less variation between the data points. However it is noteworthy that this does not definitively negate the potential impact of vent gas or operating conditions as opposed to burner type since the tests were not performed under identical conditions or with identical vent gas composition. © 2016 Callidus Technologies LLC, All Rights Reserved 3 The flow was measured across a known orifice with pressure and differential pressure readings. These pressures were in turn used to calculate the flow. Nitrogen was supplied by an onsite liquid nitrogen tank, which was vaporized atmospherically. The nitrogen flow was controlled by two separate two-inch lines, each with a two-inch ball valve and one-inch needle valve used for control. Each two-inch nitrogen line utilized separate orifice plates, pressure and differential pressure meters. The vent gas was mixed in a four-inch upstream of the flare manifold. The vent gas was burned with a single burner head having two square inches of exit area. The MP4U burner used a natural gas pilot for flare ignition, which was subsequently disconnected during operation. 3.3 Test Configuration for Program B Figure 4 - Comparison of CE for two different burner designs. 3. APPROACH AND TEST PLAN 3.1 Objective The objective of previously reported single burner testing is also presented here to establish stable operation utilizing a vent gas with less than 800 Btu/SCF NHVcz. To this end a series of tests were performed utilizing a target vent gas heat content of 500 Btu/SCF with a mixture of ethylene and nitrogen. A run length of 20 minutes was selected for each test. The vent gas pressure upstream of the burner head was varied between five and 15 Psig in order to test for the sensitivity of flame stability to vent gas exit velocity. Each run was repeated three times resulting in a total of six runs. For each test run in this low heating value vent gas test program (Program A), video recording was used to provide a record of flame stability. Vent gas composition was verified by a third party using an online gas chromatograph. Burner head pressure, manifold pressure, vent gas temperature, ethylene flow rate, nitrogen flow rate and ambient temperature were also recorded. All quantitative data aside from ambient temperature were collected continuously at an average of three-second intervals. Additional testing was performed to compare a burner of conventional type to a new alternate style of burner, which would have advantageous flame characteristics for many applications. Three burners of each type were fired simultaneously with only one being lit with a pilot burner and data was collected to demonstrate cross-lighting using low heating value vent gas. For each test run in this low heating value vent gas burner comparison test program (Program B), video recording was used to provide a record of cross-lighting and flame stability. Using partial pressures to control vent gas composition a pressurized tank was filled, which was then released into the flare manifold. Burner manifold pressure and vent gas flow rate were also recorded. All quantitative data aside from ambient temperature were collected continuously at an average of one-second intervals Butane was vaporized and supplied with an appropriate amount of nitrogen to achieve the desired vent gas heating value was supplied to a pressurized tank. The flow was measured across a known orifice with pressure and differential pressure readings. These pressures were used to calculate the flow from the tank to the flare manifold, which distributed the flare gas to three burners of the same type. The first burner in the lighting sequence on the runner used a pilot burner supplied with natural gas for flare ignition which was left in operation for the duration of the test. Radiation and flame length testing was conducted in Program B utilizing a single burner of each type. The radiation data was taken during separate test runs from those of the flame length data utilizing burners that had similar characteristics but were not of exactly the same design. In each case the alternate design is compared to the base case conventional design. 4. RESULTS 4.1 Program A: Stability Testing at 500 Btu/SCF NHVcz Stable operation of the burner flame was demonstrated on each of the six runs with an average vent gas NHVcz of 501 Btu/SCF. The resulting summary data is given in Table 1. The nitrogen-toethylene ratio was then intentionally increased to reduce the heat content until the flame became unstable and was extinguished. For both the high pressure and low pressure conditions, the flame became unstable and was extinguished at approximately 375 Btu/SCF but did not exhibit instability prior to achieving said mixture. Table 1 - Results from the Callidus MP4U low heating value burner test. For each of the critical output measures of thermal radiation -cross-lighting, flame stability and flame length -- a test plan was carried out comparing either burner type or vent gas variability. At this time a crossed experiment has not been performed for these variables. 3.2 Test Configuration for Program A Ethylene was supplied by a high pressure tube trailer, which was controlled by a two-inch ball valve and one-inch needle valve. © 2016 Callidus Technologies LLC, All Rights Reserved 4 simultaneously from multiple locations and the appropriate corrections applied. Mole Fraction of Nitrogen in Vent Gas Resulting in Unstable Flame vs Mole Weight of Flammable Components 4.4 Program B: Flame Length Results Figure 7 shows the comparison of flame lengths from test Program B for a burner of conventional design and the alternate design. Note that both the mean and the standard deviation in the flame length is less for the alternate design; not only is the flame shorter but it is more consistent in height during operation. 0.74 0.72 0.7 4.5 Program B: Cross-Lighting Results 0.68 0.66 0.64 0.62 0 10 20 30 40 50 Mole Weight of Flammable Components in Vent Gas Considered in Absence of Nitrogen 60 70 Figure 5 - The line in the chart represents the mole fraction of nitrogen in the vent gas at which a burner became unstable. The y-axis is the mole fraction of nitrogen in the vent gas. The x-axis represents the mole weight of the vent gas considered in absence of the nitrogen. 4.2 Program A: Further Analysis of Results Table 2 shows a comparison of flame stability from the Honeywell UOP Callidus conventional burner versus another type of burner. Utilizing NHVcz and LFLcz one cannot discriminate whether the flame will be stable. The Honeywell UOP Callidus test program initially utilized mixtures of ethylene and nitrogen while the other test program utilized mixtures of propylene and nitrogen. ‘Program A' was extended with further testing for both propylene/nitrogen mixtures and butane/nitrogen mixtures. Based on an extremely limited data set, a chart combining the nitrogen instability points is given in Figure 5. In this chart a trend is evident between the mole weight of the flammable components of the vent gas considered in isolation from the inert components and the ultimate mole fraction of nitrogen in the inert gas at which the burner became unstable. This chart combines data from multiple different manufacturers and from multiple different test programs. There are numerous ways to exchange the x-axis of Figure 5 with other reasonable independent variables but the general trend remains the same: the heavier (higher heat content) the flammable component, the lower inert content the burner can tolerate and retain a stable flame. This may run counter to conventional wisdom that higher heating value gas is burns ‘hotter' or is easier to burn. 4.3 Program B: Thermal Radiation Results A histogram of normalized fraction of heat release from the flame realized as radiation at grade is given in Figure 6. All radiation data is normalized to the mean of the radiation emitted by the flame of the conventional Honeywell UOP Callidus burner. The new, alternate, burner design is shown to have 15% lower radiation with 95% confidence at the same location when compared to the conventional burner. It is important to note that in the near field of a multipoint flare burner. the thermal radiation is anisotropic and the often-used spherical propagation assumption breaks down. In order to fully characterize burner performance, measurements must be taken It was determined during the testing of three burner heads for cross-lighting ability that the new burner of alternate design was capable of cross-lighting with vent gas heating values that are 85% of those for which the previous conventional generation of burner was capable. The specific inert and hydrocarbon are also likely important as was shown in the stability testing but this has not yet been fully explored. 5. THE IMPORTANCE OF ANCILLARY EQUIPMENT It has been demonstrated that the specifics of burner head design can have a large impact on the flame performance within an MPGF, and while much attention is given the burner and flare fence technology, the supporting equipment plays a vital role in the success of an MPGF installation. As such, the following practical considerations should be made. Table 2 - Comparison of stable operating conditions for different burner styles. Burner Type MP4U MP4U MP4U MP4U MP4U MP4U Other Other Other Other Stable Flame? TRUE TRUE TRUE TRUE TRUE TRUE FALSE FALSE FALSE FALSE NHVcz LFLcz 481 511 508 499 497 509 595 589 746 650 8.44 7.95 7.99 8.14 8.16 7.97 7.6 7.8 6.6 7.6 Histogram of Normalized Fraction Heat Normal Bu rn er Type Altern ate Con ven tion al 2.0 Mean StDev N 0.7597 0.2773 201 1 0.4272 212 1.5 Density Mole Fraction of Nitrogen in Vent Gas Resulting in Unstable Flame 0.76 1.0 0.5 0.0 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 Normalized Fraction H eat Figure 6 - Normalized heat release from the flare in the form of radiation. Normalization is to the mean radiation of the ‘conventional' burner tip. © 2016 Callidus Technologies LLC, All Rights Reserved 5 Histogram of Normalize Flame Length Normal 20 Bu rn er Type Altern ate Con ven tion al Mean StDev N 0.4685 0.02279 10 1 0.04828 10 Density 15 10 5 0 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Normalize Flame Length Figure 7 - Comparison of flame lengths from a conventional MPGF burner and the alternate burner of new design. Liquid carryover through the knockout drum can be extremely detrimental to flare performance. Hydrocarbon liquid droplets that exit through the flare burners can result in significant heat damage, particularly in ground flares. These same droplets can increase the smoke production from the flare and can also present a safety issue when ignited in the form of personnel hazard and equipment damage around the flare. If the droplets are not ignited, those droplets also can present a contamination issue to the surrounding environment. Given all of these negative consequences, it should be considered that on balance a knockout drum pump is relatively inexpensive. The cost of an oversized knockout drum pumping system is a fraction of the cost of repairing a ground flare due to damage. One should consider larger pumps, particularly in ground flare service. It should also be considered that these same pumps must operate during a total power failure which is often when the flare is heavily loaded. Finally the knockout drum should be located near the flare system, particularly with condensing hydrocarbons. If not the liquid in the vent gas stream may condense in the waste gas header and ultimately flow through the flare. 6. CONCLUSIONS The design of the flare burner head can have significant impact on the performance of the flare system. The performance of an MPGF system cannot be characterized by a single generic parameter such as the LFLcz or NHVcz for DRE, combustion efficiency, thermal radiation, cross-lighting ability, flame stability, or required fence height due to flame length. It does appear that among well-designed MPGF burners it is plausible that if the burner is lit and stable then the EPA required DRE is always obtained based on the vent gas that has been tested to date. It is also evident that flame stability is a function of the inert content of the vent gas taken in concert with the mole weight of the flammable constituents contained in the vent gas stream as opposed to a single parameter of heating value. However, once the acceptable vent gas composition is established for a given burner head/vent gas stream combination, vent gas heating value might be used as an acceptable proxy for determining whether a vent gas will produce an unstable flame. It does appear that for pressure assisted nozzle mix burners of proper design, some generalization may be possible across burner types based on the inert content and mole weight of the flammable portion of the gas. The crosslighting ability of a burner can likely be characterized in a similar manner but this has not been demonstrated. 7. REFERENCES [1] Author not listed, ‘Parameters for Properly Designed and Operated Flares', Report for Flare Review Panel, Prepared by U.S. EPA Office of Air Quality Planning and Standard, 2012 [2] ‘Receipt of Approval Requests for the Operation of PressureAssisted Multi-Point Ground Flare Technology'. Federal Register / Vol. 80, No. 30, 2015 [3] ‘Performance Test of Steam-Assisted and Pressure Assisted Ground Flare Burners with Passive FTIR - Garyville', EPA Document EPA-QA-OAR-2014-0738-0003, Originally 2013 referenced in 2015 [4] ‘Dow Chemical Test Report', EPA Document EPA-HQOAR-2014-0738-0008, Originally 2013 referenced in 2015 [5] Bouchard, Andrew, ‘Review of Available Test Data on Multipoint Ground Flares', EPA Document EPA-HQ-OAR2014-0738-0002, 2015 [6] Martin, Matthew et al, ‘Physical Testing of a Multipoint Ground Flare Burner Utilizing Low BTU Flare Gas', Presented at the AFRC 2015 Industrial Combustion Symposium, September 9-11, 2015 © 2016 Callidus Technologies LLC, All Rights Reserved 6 8. Appendix I - Data Table Test Condition PA1 PA1 PA1 PA1 PA1 PA1 PA1 PA2 PA2 PA2 PA2 PA2 PA2 P1H1 P1H2 P1H3 P1L1 P1L2 P1L3 P2H1 P2H1B P2H2 P2H3 P2L1 P2L1b P2L2 P2L3 P3H1 P3H2 P3H3 P3L1b P3L2 P3L3 C1 C2 C3 C4 C5 C6 C7 Stable Flame? YES YES YES YES NO YES NO YES YES YES YES YES YES YES YES YES YES YES YES NO YES YES YES NO YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES NHV 906 915 917 591 595 607 589 920 906 903 611 600 600 2170 2117 2149 2133 2134 2132 746 771 788 770 650 699 717 718 720 712 715 706 705 721 1499 1190 1499 1499 1499 1098 1499 LFL 5 4.9 4.9 7.7 7.6 7.6 7.8 4.9 4.9 5 7.4 7.8 7.8 2.4 2.4 2.4 2.4 2.4 2.4 6.6 6.3 6.3 6.3 7.6 6.9 7 7 6.9 6.9 6.9 7 7 6.9 2.75 3.46 2.75 2.75 2.75 3.76 2.75 CE 98.7 100 100 98.5 99.6 99.8 100 98.9 98.8 100 99.8 99.4 99.3 100 99.8 99.8 99.7 99.9 99.5 99.98 99.9 99.8 99.8 99.74 99.7 99.8 99.6 99.8 99.8 99.6 99.9 99.8 99.7 99.96 99.94 99.95 99.97 99.72 99.82 99.99 N2 0.016 0.012 0.013 0.365 0.354 0.356 0.372 0.016 0.016 0.016 0.337 0.376 0.377 0 0 0 0 0 0 0.634 0.62 0.621 0.621 0.684 0.65 0.655 0.658 0.524 0.526 0.523 0.531 0.533 0.524 0 0 0 0 0 0 0 CO2 0.006 0.011 0.008 0.003 0.006 0.004 0.004 0.006 0.006 0.007 0.005 0.004 0.005 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 CH4 0.937 0.931 0.935 0.605 0.617 0.612 0.595 0.935 0.933 0.935 0.633 0.592 0.591 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.243 0.246 0.249 0.241 0.239 0.242 0 0 0 0 0 0 0 C2H4 © 2016 Callidus Technologies LLC, All Rights Reserved 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 C2H6 0.038 0.042 0.039 0.025 0.021 0.026 0.027 0.039 0.04 0.037 0.022 0.026 0.026 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C3H8 0.003 0.004 0.004 0.002 0.002 0.002 0.002 0.004 0.004 0.004 0.003 0.002 0.002 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C4H6 0 0 0.001 0 0 0 0 0.001 0.001 0.001 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C3H6 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0.366 0.38 0.379 0.379 0.316 0.35 0.345 0.342 0.233 0.228 0.229 0.227 0.228 0.234 0 0 0 0 0 0 0 7