You are on page 1of 31

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Summary This document details a thermodynamically sound method for calculation of flare tip pressure. The pressure of a flare tip is an important parameter in predicting flare operational performance. The justification for this document is because standard reference document1 API-RP-521 entirely omits details for determination of flare tip pressure. Another standard reference document2 GPSA Engineering Data Book uses Straitz method a to l to determine flare tip pressure. The Straitz method determines pressure drop as 1.5 velocity heads. However Straitz method uses a form of velocity head strictly suitable to liquids. The Straitz method is not suitable where pressure change is greater than 10% of initial pressure, (Crane4 pg. 3-3). The approximate limits of Straitz method are between 0.25 Mach and 0.4 Mach for methane and propane respectively. Another recently published3 pressure drop method is similar to the Straitz method. This paper concurs the recommendation Crane TP-410: gas flow across a nozzle is adiabatic (Crane4 pg. 1-9). The adiabatic pressure drop calculation has been applied successfully to a wide range of flow nozzles such as gas well chokes, metering orifices, and relief/ control valves. Flare Tip Pressure Perhaps pressure of the flare tip is the most critical parameter in evaluation and design of flares. This is due to flare tip sizing parameters relate to tip velocity and tip velocity relates to pressure. The tip pressure sets the pressure profile through out the entire flare header. Ignoring Flare tip pressure drop will result in under estimation of noise level by between 7 and 13 decibels. Additionally ignoring flare tip pressure will under rate the flow capacity of subsonic flare. However most discussions on flare sizing omit details on calculation of tip pressure or relegate tip pressure to a set of ranges8. Some methods t, u relates tip velocity to the radiant heat11 fraction emitted by a flame. The proposed method expresses pressure drop as an implicit function, based on the adiabatic gas flow equation.

This method is generally applicable to a wide variety of nozzle flows, such as orifice plates, relief valves at subsonic flow, control valves, etc. The details of this equation are presented in the following Table, based on common field units, as applied to gas well choke nozzles. Since a flare tip is an engineered piece of equipment just like a choke valve, basic performance parameters should therefore be specifiable. A calculation spreadsheet was prepared to allow pressure drop calculation per the adiabatic method. The solver routine was implemented since pressure drop is an implicit value in equation 5.2, above. The Crane method4 is used to generate the initial estimate for the implicit solution.

[1]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

The Choke Valve Method9


Mach number is a ratio between sound velocity in gas divided into gas flow velocity. Determination of Critical Pressure Drop or Critical Pressure Ratio, CPR

The CPR is the minimum pressure ratio for sonic flow, and for any larger pressure ratio, flow inside a nozzle is subsonic, M<1. The pressure ratio is conventionally defined as pressure upstream divided into downstream pressure. Since flow is always in a direction of decreasing pressure, pressure upstream will be larger than pressure downstream. At pressure ratio of 1, flow is zero. From thermodynamics CPR is determined by formula 5.1. d/D = 33/36 = 0.92

C = 0.92 + 0.33 + 0.09 = 1.34 Since Nre>1E6, use C =1.2, as per example

Nre = 20 (651,430)*0.9883/(0.01245*33) =3.2E7

[2]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

The following charts offer a comparison of spreadsheet results using the Zink Company publication5. A certain amount of variation is expected, given the lack of accurate flare tip pressure from most flare tip performance calculation methods.

[3]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer References:


1.

Anon., Guide for Pressure-Relieving and De-pressuring Systems, RP-521, 5th edition, The American Petroleum Institute (API), Washington, D. C., January 2007 & update ISO 23251, {no change to procedures offers no specific directives to
calculate pressure drop and pressure drop is required to determine tip pressure, which is required to determine Velocity or Mach Number} : With a proper tip design, most commercial flares have flame retainers that restrict flow area by 2 percent to 10 percent; the flame of the main stream can be anchored in the boundary regions where velocity gradient would otherwise far exceed the critical value for blowoff . There is evidence; 16, 17, 18; that flame stability can be maintained at relatively high velocities depending on the discharge properties and the type of tip used. Both blowoff and flashback velocities are greater for fuels that have high burning velocities. Small amounts of hydrogen in a hydrocarbon fuel widen the stability range because blowoff velocity increases much faster than flashback velocity. Designs may be based on velocities of 0.5 Mach or higher, if pressure drop, noise, and other factors permits. Flare stack diameter is generally sized on a velocity basis, although pressure drop should be checked. One may want to permit a velocity of up to 0.5 Mach for a peak, short-term, infrequent flow, with 0.2 Mach maintained for the more normal and possibly more frequent conditions for low-pressure flares, depending on the following: (a) volume ratio of maximum conceivable flare flow to anticipated average flare flow; (b) the probable timing, frequency, and duration of those flows; and (c) the design criteria established for the project to stabilize flare burning. 5.4.3.1.1 However, sonic velocity operation may be appropriate for high-pressure flares. Smokeless flares should be sized for the conditions under which they are to operate smokeless. Equation 23 or 24 can be used to calculate the Mach number (see 5.4.1.3.2). Velocity limitations imposed by CFR 60.18 do not apply to flares in emergency relief service. Pressure drops as large as 2 psi (14 Kpa) have been satisfactorily used at the flare tip. Too low a tip velocity can cause heat and corrosion damage..//offers no specific directives to calculate pressure drop//

2.
3.

Anon., Engineering Data Book, 10th edition, Section 5 - Relief Systems, Gas Processors Suppliers Association, 1987. Sizing based on Straitz method, see below citations A to L Anon, Quick Estimate of Flare Tip Pressure Drop, Chemical & Process Technology pp11-14, March 2009 AFSA or

FLARENET is commonly used for flare network modeling. In calculating back pressure at the PRV, flare tip pressure drop is required. However, the pressure drop of flare is subject to Flare tip vendor design. How shall engineer determine the pressure drop of flare tip without vendor information especially during conceptual design ? Pressure drop provided by vendor shall always be used during detailed design For a sonic flare tip, pressure drop may range: 3 - 5 bar. (or) .. up to 7 bar . For a subsonic flare tip, pressure drop is generally very small (possibly lower than 1 bar) equation offered: dKpa= 10 ^ {a(Log (scm/hr/3.6))+b}., a<1.5,2>, b<-8.3,-5.3>, standard definition is 101.325 kPa abs and 15 deg C DN250 tip, dP = 10 ^ ( 1.77 x Log Q - 5.29) = 10 ^ ( 1.77 x Log 30000/3.6 5.29) = 46 kPa //note; this is not adiabatic pressure drop, Crane TP-410 states: flow across nozzle is an adiabatic process//

4.

Anon, Flow of Fluids Through Valves, Fittings, and Pipe, Technical Paper Number 410, Crane Co., New York, 1957, p1-9 compressible fluids discharge considered to be adiabatic.. supported by experimental data discharging air to
atmosphere pp. 3-3 compressible flow when dP less than 10% of P12-14&15, and A-19 to A-21 detail Y method for gas flow

5.

Schwartz, R. E. and White, J. W. Flare Radiation Prediction: A Critical Review ZINK Corp., AIChE 1996 Annual LP Symp. Session 12 Flare Stacks The authors have designed hundreds of flares which have discharge velocities greater than 0.5

6.
7.

Mach at maximum flow. . In choosing a discharge velocity one must always take care to avoid over pressuring the relief system. // offers no calc guidance except refer to API//

Beychok, M.R. Fundamentals of Stack Gas Dispersion, Ch11 Flare Stack Plume Rise, 4thed. New Port Beach CA. 2005 Vatavuk, B., Pollution Control Equipment Cost, U.S. EPA, OAQPS Ch7 Flares Radian Corp RTP NC 1998 discuss

need to know velocity and pressure at tip but offers no calc guidance) (7.1) The EPA requirements for flares used to comply with EPA air emission standards are specified in 40 CFR Section 60.18. The requirements are for steam-assisted, air-assisted, and non-assisted flares. Requirements for steam-assisted, elevated flares state that the flare shall be designed for and operated with: C an exit velocity at the flare tip of less than 60 ft/sec for 300 Btu/scf gas streams and less than 400 ft/sec for >1,000 Btu/scf gas streams. For gas streams between 300-1,000 Btu/scf the maximum permitted velocity (V, ft/sec) is determined by the following max equation: log10(Vmax)=((Bv+1214)/852), where B is the net heating value in Btu/scf, v, The actual maximum capacity of a flare tip is usually limited by the vent stream pressure available to overcome the system pressure drop. Elevated flares diameters are normally sized to provide vapor velocities at maximum throughput of about 50 percent of the sonic velocity of the gas subject to the constraints of CFR 60.18. The physical limitation on the quantity of steam that can be delivered and injected into the flare flame determines the smokeless capacity of the flare. Smokeless capacity refers to the volume of gas that can be combusted in a flare without smoke generation. The smokeless capacity is usually less than the stable flame capacity of the burner tip. A pressure drop calculation is required at this point to ensure that the vent stream has sufficient pressure to overcome the pressure drop occurring through the flare system at maximum flow conditions. The pressure drop calculation is site specific but must take into account losses through the collection header and piping, the knock-out drum, the liquid seal, the flare stack, the gas seal, and finally the flare tip. Piping size should be assumed equal to the flare tip diameter. Schedule 40 carbon steel pipe is typically used. The design pressure drop through the flare tip can range from 0.1 to 2 psi with the following approximate pressure drop relationships:[5] Gas seal: 1 to 3 times flare tip pressure drop Stack: 0.25 to 2 times flare tip pressure drop Liquid seal and Knock- 1 to 1.5 times flare tip pressure drop plus out drum: pressure drop due to liquid depth in the seal, which is normally 0.2 to 1.5 psi. Gas collection system: calculated based on diameter, length, and flow. System is sized by designer to utilize the pressure drop available and still leave a pressure at the stack base of between 2 and 10 psi. Typical total system pressure drop ranges from about 1 to 25 psi.[5] This approach will encourage higher flare tip exit velocities, which promote higher combustion efficiencies..

[4]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer


8. Czaikowski, M., P. Process Engineering Design Guide Relief System Design: 3DGV88D1, Bechtel Corp. 2004
11.5 Super-sonic Furnace Burner used offshore to save weight. 11.5.3.1 The tip diameter and pressure drop through the stack, for elevated flares, should be checked and confirmed by the flare vendor to ensure that excessive back pressure has not been imposed in the header system. 11.3 For initial estimating purposes the following guidelines may be used: Equipment Pressure Drop (psi) Flare Tip 0.5 2.0 Molecular Seals 0.5 1.0 Fluidic Seals 0.1 Seal Pots 1.0 1.5 Note: For ground flares, additional pressure drop should be allowed for staging valve operation. The staging control valve differential pressure is usually 5 - 10 psi or more. The ground flare vendor should be consulted to confirm the pressure drop. Note that the API method (Example 1 in API 521 Appendix C) provides satisfactory results when the target is relatively far from the flare tip, but overestimates the radiation when the target is close by (meaning API not good means for determination of flare stack height: opa). Section 3.19 should be followed. This section should be interpreted to read that the depressuring target of 50% of design pressure or 100 psig (whichever is lower) in 15 minutes should be used whenever leakage of vessel contents leading to a fire is a concern. A standard of 50% of design pressure in 15 minutes is used when the depressuring is for process reasons only, not fire, and the vessel wall thickness is 1 or greater. John Zink Co. suggests a 0.01 ft/sec minimum purge gas rate under most conditions, with up to 0.03 ft/sec minimum in special circumstances. In general, the purge gas rate should be specified by the molecular seal vendor. vendor of the device should confirm the steam requirements for smokeless burning. Most often it is impossible or economically impractical to supply sufficient steam to render the flare smokeless under maximum relieving conditions. ergo.

9. 10.

Gao, B. & Ghalambor, A. Natural Gas Engineering Handbook, Ch7 Choke Performance GPC Houston TX 2005
This reference provided the adiabatic pressure drop equations recommended for Tip Pressure Drop calculation, this equation was selected as it is in a field units format. The examples are direct cuts from this document.

Ludwig, E.E., Applied Process Design For Chemical and Petrochemical Plants, Vol. 1, 3rd Gulf Publishing Co., Houston, TX, Section 9.125-9.129 Flares/Flare Stacks 1995. {Offers no calc method on flare pressure drop, most material offered
follows GPSA (Straitz Method)or API methods: 9.127: For specific details, consult a flare system design manufacturer (who use proprietary systems?), 9.126: Pressure drops across the tip of the flare have been used satisfactorily up to 2psi. It is important not to be too low and get flashback (without a molecular seal) or blow off where the flame blows off the tip; 9.137, V, fpsexit, = 550((dpinchH2O)/55), (the method leads to the a quadratic tip pressure as density depends on pressure}

11. Guidard, S.E., Kindzierski, W.B. and Harper N.,. Heat Radiation from Flares. Report prepared for Science and Technology Branch, Alberta Environment, ISBN 0-7785-1188-X, Edmonton, Alberta, 2000 A matrix summarizing
which parameters have been used to determine the fraction of heat radiated for each of these relationships is detailed. The applicability of these relationships to the general case is limited. The theoretical or empirical conditions for which many of these relationships are based upon are situation-specific. In addition, limited information was provided in many instances on numerous parameters that are known to influence flare heat radiation losses (e.g. stack exit velocity, crosswind velocity, aerodynamics of the flame, etc.) .. Determination of the thermal radiation emitted from flares is important in facility design, since it establishes the required flare site and stack height in order that workers and equipment are protected Chamberlain (1987) noted that Equation 8 (Ludwig60/API Heat Rad. Eqn.) has been successfully applied to onshore refinery flares for many years. However, he indicated that it is of limited use offshore because it can only predict thermal radiation accurately in the far field (the opposite to what Brzustowski and Sommer (1973) reported).
method on flare pressure drop, most material offered follows API methods or based on Kents method:}

12. Banerjee, K. Flare Gas Systems Pocket Handbook 1st Ed. Gulf Publishing Co., Houston, TX, 1985 {Offers no calc

13. Zelensky M.J. Sour Well Test Flaring Permit Application Process and Dispersion Modeling Nonographs, Alberta Energy Utilization Board, Calgary 09/2001 4.3.2 The minimum recommended exit velocity at the flare tip is based on the flare
research at the University of Alberta (1998). They found that the flare efficiency was strongly sensitive to crosswinds and that higher exi t velocity flares are less prone to becoming inefficient. In general, flare efficiency exceeded 99 percent if the flare tip exit velocity exceeded the wind speed at the flare tip. The wind speed at the flare tip changes hour-by-hour while flaring so a statistical approach was taken. The 99th percentile wind speed (U99, m/s) for Alberta was determined to be 10.28 m/s at the standard 10 m reference height, based on AENVs six regional meteorological data sets. U99 at 10 m is adjusted to the wind speed at flare tip height (h, m) using a power profile with an exponent of 0.15 for the neutral stabi lity conditions associated with high wind speeds. The approach adopted by the U.S. EPA (1992b) from Beychok (1979) was used to account for the flame height required for combustion. The effective stack height (Hs, m) is based on the flare stack tip height (H, m) and a portion of the flame lengt h (Hf, m) that depends on the maximum heat release rate (EM). Hf = 0.00456*(EM)0.478 Hs = Hf+h The flame length correlation is based on the maximum heat release rate assuming complete combustion. The constant of 0.00456 assumes the flame is tilted at an angle of 45 degrees to the horizontal by the wind and that the effective stack height is the center of the tilted flame. Using one-half of the flame length is a conservative assumption. High wind speeds are required to tilt the flame 45 degrees. This approach is a reasonable compromise and avoids the complications of making the effective height a function of wind speed and actual flare tip exit velocity

[5]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Citation References as critical in the developmental concepts of flaring technology


a. b. c. d. e. f. g. h. i. j.

Straitz, J.F., and R.J. Altube, Flares: Design and Operation, Pub. National Air-Oil Burner Co., Inc. 1980 Straitz, J.F., Make the Flare Protect the Environment, Hydrocarbon-Processing, (56), pp131/35, Oct 1977 Straitz, J.F., "Flaring for Safety and Environmental Protection" Drilling-DCW, November, 45-48. (1977) Straitz, J.F., "Solving Flare Noise Problems" Inter. Noise 78, San Fran. 8-10 May pp1-6 (1978) Straitz, J. F. "Nomograph Determines Proper Flare-Stack Height." Oil, Gas, & Petrochem Equipment, v.25n.10, p.25
August, 1979 (ISSN 00301353)

Straitz, J. F. "Smokeless Flaring at High Rates." ASME Pet Mech Eng Symp, Philadelphia, PA, pp.105/11, Sept.1982 Straitz, J. F. High Performance Offshore Flares." 4th International flare seminar Norway, 1986 Straitz, J.F., "Flare Technology Safety" Chem. Eng. Progress, v.83n.7 July, pp53-62, (1987) Straitz, J.F., "Improve Flare Design" Hydrocarbon Processing, v.73n.10 October, pp61-66, (1994) Straitz, J.F., et-al., "Flare Pilot Design" 1996 International Symposium, Combustion in IndustryStatus and Needs into the 21st Century, Baltimore, MD 1996 k. Straitz, J.F., "Clearing the Air, About Flare Systems" Chemical Engineering, 103(9) 1996 l. Straitz, J.F.,"Improve Flare Safety to meet ISO-9000 Standards" Hydrocarbon-Processing. v.75n.6,4pp Jun 1996, m. Brzustowski, T.A. and Sommer E.C., Predicting Radiant Heating from Flames, Proceedings Div. of Refining, API v. 53, pp. 865-893, 1973. n. Brzustowski, T. A., A Model for Predicting the Shapes and Lengths of Turbulent Diffusion Flames Over Elevated Industrial Flares, Proceedings 22nd Canadian Chemical Engineering Conference, Toronto, 1972. o. Oenbring, P.R. and Sifferman T.R., Flare Design: Are Current Methods Too Conservative? HydrocarbonProcessing, v59n5, pp124-129, 1980 p. Kent, G. R., Practical Design of Flare Stacks Hydrocarbon Processing, v43n8 pp121-125, 1964 q. Hajek J. D. and Ludwig E. E., How to Design Safe Flare Stacks, Part 1, Petro/Chem. Engineer; 1960, v32, n6, pp. C31-C38; Part 2, Petro/Chem. Engineer; 1960, v32n7, pp. C44-C51, 1960 r. Narasimhan, N.D., Predict Flare Noise, Hydrocarbon-Processing, v.65, n4. p133, 1986 s. Tan, S. H., Flare System Design Simplified, Hydrocarbon-Processing , v46n1, p172-176, 1967 t. Chamberlain, G. A. Developments in Design Methods for Predicting Thermal Radiation from Flares, Chemical Engineering, Research and Design, Volume 65, July 1987 u. Cook, D. K., Fairweather, M., Hammonds, J. and Hughes, D. J. Size and Radiative Characteristics of Natural Gas Flares. Part 1 Field Scale Experiments, Part 2 Empirical Model, Chemical Engineering, Research and Design, Volume 65, July 1987, pp318-325, 310-317., 1987 method for rad of sonic flares v. Barnwell, J. & Marshall, B. K. Offshore Flare Design To Save Weight American Institute of Chemical Engineers Meeting, November 1984, San Francisco, California, 1984 w. Smith-SK; Selle-GK, Supersonic, High Pressure, Low Radiation Flare System Design, Offshore-TechnologyConference,-Annual-Proceedings. v 4 1997, Offshore Tech Conf, Richardson, TX, USA. 14p Proceedings of the 1997 29th Annual Offshore Technology Conference, OTC'97. part 4 of 4. Houston, TX, USA x. Magda-W; Marcinkowski-T; Mazurkiewicz-B-K Cantilevered flame boom the effect of wind on flare exit angle Proceedings of the Sixth (1987) International Offshore Mechanics and Arctic Engineering Symposium. Houston, TX, USA Publ by ASME, New York, NY, USA p 275-279 y. Bjorge T, Bratseth A, Measurement of radiation heat flux from large scale flares JOURNAL OF HAZARDOUS MATERIALS 46: (2-3) 159-168 APR 1996 z. Beeri Z, Blunsdon CA, Malalasekera WMG, Dent JC Comprehensive modeling of turbulent flames with the coherent flame-sheet model .2. High-momentum reactive jets JOURNAL OF ENERGY RESOURCES TECHNOLOGY-TRANSACTIONS OF THE ASME 118: (1) 72-76 MAR 1996

[6]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

F Factor Use 0.15 Hydrogen 0.2 Methane 0.3 Higher MW Hydrocarbons The following thermal radiation limits apply for personnel: Conditions Maximum Heat Intensity BTU/hr/
sf Operating areas where personnel are not shielded from radiation 1000 Operation areas where

personnel are shielded from radiation, by virtue of equipment, pipeways and the like 1500 Operating areas where no shelter exists and only escape is required 3000 Operating area in which personnel normally do not work and where shelter is available 5000 The issue of whether or not these limits

include solar radiation is left open in APIRP 521. The use of a molecular seal (See Section 11.6.2) should allow substantial reduction of the purge gas rate. John Zink Co. suggests a 0.01 ft/sec minimum purge gas rate under most conditions, with up to 0.03 ft/sec minimum in special circumstances. In general, the purge gas rate should be specified by the molecular seal vendor. Manufacturer of the device should confirm the steam requirements for smokeless burning. Most often it is impossible or economically impractical to supply sufficient steam to render the flare smokeless under maximum relieving conditions. The following guidelines are offered as reasonable compromises for smokeless flare design basis: Design for the maximum individual requirement, such as the FCC relief gas, that is normally large and contains large quantities of unsaturates. Design for 20% of maximum rate that typically results in about 95% of the relief occurrences being smokeless. Design for maximum rate but provide remote (adjustment from control house) control of steam addition. Governing regulations or permits should always be consulted for verification of smokeless operation requirements. In cases where specific direction or data is not available, the practice outlined in API RP 521

The basic formula used to calculate the flows discharging from ruptured tubes is: W = 1891 K Y d2 (P 1)0.5 Where: W = Flow, lb/hr at sonic flow conditions K = Discharge Coefficient Y = Expansion Factor d = Tube ID, in P = Pressure drop, psi 1 = Upstream density, lb/ft3 This is the formula for flow through nozzles and orifices given in Crane Technical Paper 410 Equation 3-21 and 3-22. To account for flow from both ends of the ruptured tube, the calculated flow rate must be multiplied by a factor of two. In calculating discharge due to tube rupture, assume that both halves of the break behave as square edged orifices with K = 0.7. The expansion factor Y is unity for a liquid and is given by the figure in Appendix H for a gas or by the equation: Y = 1 ( 0.41 / n ) ( P / P1 ) where: Y = Expansion Factor n = k = Cp/Cv for an ideal gas P = Pressure drop, psi P1 = Upstream Pressure, psia
log10(Vmax)
[7]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

This minimum velocity is dependent on both gas composition and diameter and can range from insignificant amounts on small flares to 0.5 ft/sec on greater than 60-inch diameter units.[5] The average pilot gas consumption based on an energy-efficient model is 70 scf/hr (of typical 1000 Btu per scf gas) per pilot burner.[6, 7, 8, 9, 10] The number of pilot burners, N, based on flare size is:[6, 7, 8, 9, 10] 60=4pilot
4.3.1 Minimum Flare Tip Diameter The minimum recommended flare tip diameter (DMIN, mm) is based on the maximum flaring rate (QMAX, 103m3/d) divided by the maximum recommended exit velocity (VMAX, m/s). It is set to prevent flame lift-off with the following equations: DMIN = 1000 mm/m*{(4*QMAX*103 m3/103m3)/(24 h/d*3600 s/h**VMAX)}1/2 VMAX = MIN {122, 10[(NHV+28.8)/31.7]} MIN is the minimum function. EUB Guide 60 references a U.S. EPA relationship (U.S. CFR Part 60) that specifies the maximum exit velocity (VMAX) as a function of the net heating value (NHV, MJ/m3). Higher heating value flare gas can maintain a stable flame at higher exit velocities than lower heating value flare gas. The upper limit of 122 m/s corresponds to a heating value of 51.1 MJ/m3, and is approximately a Mach number of 0.3 (i.e. 30% of the speed of sound). The U.S. EPA equation is consistent with the rule of thumb for flare tip design to not exceed a Mach number of 0.2 for continuous flares and 0.5 for emergency flares 4.3.2 Maximum Flare Tip Diameter The maximum recommended flare tip diameter (DMAX, mm) is based on a representative average flaring rate of one-half of the maximum flaring rate (0.5*QMAX, 103m3/d) divided by the minimum recommended exit velocity (VMIN, m/s) at the flare tip. It is set to avoid flame downwash with the following equation: DMAX = 1000 mm/m*{(4*0.5*QMAX*103 m3/103m3)/(24 h/d*3600 s/h**VMIN)}1/2 VMIN = U99*(h/10)0.15 DMAX is provided for guidance in sizing the flare tip and does not effect the dispersion modeling. VMIN is used in the flare tip downwash adjustment described in the next section. The representative average flaring rate used in the calculation avoids having the maximum recommended diameter less than the minimum recommended diameter. flaring rate is much less than the one half of the requested maximum flaring rate, then the flare tip should be sized closer to the minimum recommended diameter. 4.5 Ground Level Radiation EUB Guide 60 specifies that flares must be designed so that the maximum radiant heat intensity at ground level will not exceed 1500BTU/hr/sf. The following equation is based on the GPSA (1998) and API (1997) procedures referenced in EUB Guide 60, but are simplified to result in conservative predictions. The ground level thermal radiation intensity depends on the maximum flaring rate (QMAX,), net heating value (NHV), radiation loss (RAD, fraction), and the effective stack height (Hs, m), and is given by: IR = (QMAX*NHV*RAD)/(4**Hs2) Both referenced procedures require the background thermal radiation to be included in the total radiation calculation. A maximum background thermal radiation of 1.04 kW/m2, the upper range of maximum values according to the GPSA (1998), is added to the predicted value to compare to the guideline. Maximum background thermal radiation values in Alberta would be less. For sour gas flares, ground level radiation is not a limiting factor in the selection of the stack height; the dispersion of SO2 determines the required height. A 12 m flare handling about 569 103m3/d of methane equivalent heating value gas meets the thermal radiation guideline. An 18.3 m flare can handle about 1206 103m3/d and meet the guideline. Where C is the velocity of sound (m/s), L the sound level (dB), L30 the sound level at 30 m (dB), M the mass flow (kg/s) and PR the ratio of the upstream to the downstream pressure. For L in the function A12.4.2 API RP 521 gives a correlation with linear scale for L and logarithmic scale for PR. The correlation is two straight lines with a break point at the two lines being defined by the additional points. For other distances the sound level may be Lp is the sound level (dB) and r the distance from the stack tip (m). Green states that emergency relief vents tend to be the most powerful noise sources on chemical plant. The noise from a jet is greatest in the 60 cone about the axis and hence a vertical orientation is usually to be desired. He indicates that a vent should be so located that the sound level does not exceed 125 dB in areas where operators might be, however infrequently. The jet noise is proportional to the sixth to eighth power of the exit velocity. Methods of reducing noise include velocity reduction and silencing, but reduction of velocity is in direct conflict with the need for a high velocity to promote dispersion. 5.9.2 EPA Requirements EPA (1988) published criteria for flare operation (40 CFR 60.18) which apply to normal process flare systems as follows. (Even though smokeless is not normally required for emergency flares, this information is included for reference purposes since local regulations may require smokeless flaring.) Steam-assisted or air-assisted flares require the gas to have a minimum net heating value of 300 Btu/SCF. Nonassisted flares require the gas to have a minimum net heating value of 200 Btu/SCF. No visible emissions. A five minute exception period is allowed during any two consecutive hours. A flame must be present at all times when emissions may be vented. The presence of a pilot flame shall be monitored using a thermocouple or equivalent device. Maximum tip velocity is defined as a function of net heating value of the feed to the flare in Table 5.3 The equation for maximum tip velocity for heating values between 300 and 1000 Btu/SCF: (5.9.1) where Vmax is the maximum permitted velocity at flare tip, ft/s, and Bv is the net heating value of feed to flare, Btu/SCF. TABLE 5.3 Maximum Tip Velocity Heating Value of Gas, Btu/SCF Maximum Tip Velocity, ft/s . <300 60 300-1000 See Eq. (5.9.1) >1000 400 Stone et al. (1992) state that it is standard practice to size the flare tip so that the design velocity is 80% of Vmax. Too low a tip velocity can cause heatand corrosion damage. The equation for minimum tip diameter is then (5.9.2) where Dmin= minimum tip diameter, inches Qtot = total volumetric flow rate, actual CFM Vmax = maximum tip velocity, ft/s; determined from Table 5.3. The selected flare tip diameter, D, is the calculated diameter, >min, (inches), rounded up to the next commercially available size. The minimum flare size commercially available is 1 inch; larger sizes are available in 2-inch increments from 2 to 24 inches and in 6-inch increments above 24 inches. API method. API 521 recommends a maximum tip velocity corresponding to a Mach Number of up to 0.5 for peak, short-term, infrequent flow emergency discharges. The Mach Number for a given tip diameter and flow rate may be calculated from where M 2 = Mach Number at flare tip outlet W = gas flow rate, Ibm/h P pressure of gas just inside flare tip, psia D = flare tip diameter, ft Z = compressibility factor of gas under conditions at tip T = absolute temperature of the gas just inside the flare tip, 0R k = ratio of specific heats, C^/CV Mw = molecular weight of the gas. The tip diameter calculated by the API method [Eq. (5.9-3)] may be smaller than that calculated by the EPA criteria

[8]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer


(Table 5.3), and if so, the larger diameter should be selected (Shore 1995). Too low a tip velocity can result in combustion taking place inside the stackthis is called burn-back, or sometimes a flame dip. This can occur when a slow upward flow of lighter-than-air gas allows air to flow downward along the stack wall. A diffusion flame propagates down into this region and it is quenched at the wall. Air then flows downward again causing another burn-back, and the cycle is repeated again. Seebold (1984) suggests that a tip velocity of 1-3 ft/s is required to prevent flame dip. Straitz (1996) suggests that an effective air seal will prevent burn-back and burning inside the flare tip at flare burner tip velocities as low as 0.03 ft/s. The approximate limiting tip velocity is given by equation: where D is the flare tip diameter, ft, and V is the tip velocity required to prevent flame dip, ft/s. A pressure drop calculation should be made to ensure that the vent stream has sufficient pressure to overcome the pressure drop occurring through the flare system at maximum flow conditions. The pressure drop calculation is site specific but must take into account losses through the collection header and piping, the knock-out drum, the liquid seal, the flare stack, the air seal, and finally the flare tip. If sufficient pressure is not available, the economics of either a larger flare system or a gas mover such as a fan or compressor must be considered. A pressure drop as high as 2 psi across the flare tip has been used satisfactorily. This pressure drop is the result of a flame retainer at the flare tip included in most commercial flares, which restricts flow area by 2-10% (API 521). RP 521 gives a correlation with linear scale for L and logarithmic scale for PR. For other distances the sound level may be obtained from Lp where Lp is the sound level (dB) and r the distance from the stack tip (m). Green states that emergency relief vents tend to be the most powerful noise sources on chemical plant. The noise from a jet is greatest in the 60 cone about the axis and hence a vertical orientation is usually to be desired. He indicates that a vent should be so located that the sound level does not exceed 125 dB in areas where operators might be, however infrequently. The jet noise is proportional to the sixth to eighth power of the exit velocity. Methods of reducing noise include velocity reduction and silencing, but reduction of velocity is in direct conflict with the need for a high velocity to promote dispersion

[9]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Summary
Capacity of 33 inch nozzle was checked by five methods, Table 1. The resultant capacity calculated by these methods, meet or exceeded requirements (570#/sec or 2.05MM#/hr) as given in Table 2. An additional constraint is that Flare Nozzle Mach Number must not exceed 0.8 Mach number. Mach number is calculated by API-RP-521 equation 23. An additional constraint is that nozzle pressure must be less than critical pressure ratio. The critical pressure ratio is the maximum pressure ratio that is achieved prior to sonic flow conditions. This pressure ratio, PCR, is determined as given below. All methods are described in the following write up. Table 1
Method calc'd cap #/sec Tube Rupture RV 572.0 Orifice Plate 678.0 Crane TP410 K=0.5 678.0 Choke Nozzle 663.9 Mach # 684.2 P(w) TP410-b K=0.5 570.0 tip size Pup P down inch psia psia 33.0 25.90 14.7 33.0 26.40 14.7 33.0 26.40 14.7 33.0 26.00 15.0 33.0 26.40 14.7 33.0 23.90 14.7 MW #/mol 28.66 28.66 28.66 28.66 28.66 28.66 K cp/cv 1.25 1.25 1.25 1.25 1.25 1.25 T R 520 520 520 520 520 520
Ma ch # RP521.e 23

0.68 0.79 0.79 0.79 0.80 0.74

Max Pup psia 26.49 26.49 26.49 27.03 26.49 26.49

# 1 2 3 4 5 6

Table 2 HP Flare Upset Sizing Conditions, PFD Stream 215- of G80-NA-J-54992

As seen from Table 2, sufficient pressure is available at flare KO Drum to achieve a pressure necessary for flaring 570#/sec of gas, 23.9 psia. This leaves 37psia less 23.9 or about 13 psi for pipe and flare system pressure losses. The ability of a 60 inch flare header to achieve required relief rate with pressure losses at or below 13 psi is detailed in this write up. Since limiting case is for emergency depressurization, a review is made for heat radiation limits under transient flow conditions. It is pointed out that high values of transient heat radiation are permitted by API-RP-521. The heat flux requirement for a transient condition is not same as that for continuous operation.
A simplified Flare Tip sizing routine by Mach number is as follows: 1) calculate Critical Pressure and select tip pressure slightly less than calculated critical pressure, 1b) check pressure balance to verify tip pressure. 2) Use MW, mass rate, and th Temperature from PFD 3) Calculate Cp/Cv from Figure 16-2 of GPSA 9 Edition. 4) Use formula 23 of API-RP-521 to find required tip diameter for allowable Mach number, 0.80 in this instance.. The same method may be used to determine capacity of a tip. (k/(k-1)) PCR = (2/(k+1))

[10]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Overview
No small amount of deliberation has been made over whether or not an old flare, F3, will pass a mass rate of 569 pounds per second of 28.7MW hydrocarbon gas available at 37 psia from flare KO drum. F3 is a decommissioned unit. It has a 12 foot diameter base with a 36 Flare Tip (point 3) sitting atop an 82 inch molecular seal, point 1. Between the 12 foot base and the 82 inch section is another section of approximately 10 foot diameter. It is the premise given here that a 36 inch tip is adequate for the specified service. Considerations are given from a basic engineering point. Also additional witness are called to support this position, namely, Mr. Farris Relief Valve, Mr. Daniel Orifice, and Mr. Crane TP-410 method of example 4.21, Mr. Choke Valve, and Mr. ISA C. Valve. All these witness are credible by virtue of regular application. I) Basic Diagram with Energy Balance Considerations

The Bernoulli energy balance equation for above condition is: 144(P/) 2 + V22/62.4 = 144 (P/)3 + V32/64.4 + Hf EQN. 1

If is in #/CF with pressure in #/sq-in, the following conversion factor must be introduced, 144 in2/ft2, so pressure head is in feet, as is velocity head when V is ft/sec and g is 32.2 ft/s/s. The friction losses, Hf , are expressed in terms of V2. They may consist of following basic terms:

[11]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

1. 2. 3. 4. 5. 6. 7. 8.

Acceleration of gas from low velocity state to higher velocity state, Ka Turbulence losses in contraction (vena cava) section, Kv Deceleration losses as gas expands to P3, Kd And lastly a thermodynamic term, to account for adiabatic expansion density changes, Kt Collectively these terms are called+ Ki to keep algebra simple As will be shown, the Crane method just lumps all terms into a Ki to keep algebra simple Losses are expressed in terms of inlet V, since V in dispersion cone is not precisely known It will be shown for reasonable amounts of expansion this velocity head becomes insignificant and that inclusion of adiabatic expansion factor eliminates need for item 4. EQN. 2 EQN 3 a&b

144(P/) 2 + V22/62.4 = 144 (P/)3 + V32/64.4 + ( Ki) V22/64.4

Velocity may be expressed in terms of mass rate as V=(w/(A )) or V(A) =w.

Since continuity requires w2 = w3, then (VA)2 = (VA/)3 & V3 =(VA)2/(A)3 =(V)2*(d2/d3)2 /3 For any gas = PM/10.73zT, where P is psia and M lb/mol and T is degree Rankin. For adiabatic expansion density changes are related to pressure ratio raised to inverse of heat capacity ratio. The area ratio is accounted for by term beta, . Area ratio is just a ratio of diameters squared. For the velocity head term it is to the forth power. Consider situation where the diameter triples (as shown in above diagram) and initial conditions are: V2 =800fps MW=29 P2=27, T2=520R, beta =3, k=1.25, den2=27*29/10.7/520=0.14 Conditions at 3: den3 =0.14*(15/27)^(1/1.25) =0.09, V3=800*(.14/.09)*(1/3)^2=142, the Vhead3= (0.09)*142^2/64.4/144 = 0.20psi, Vhead2 = 0.14*800^2/64.4/144= 9.7psi. V head 3 is less than 2% of initial V head and only 3% of exit pressure head. Thus for any reasonable expansion, velocity head on Right hand side is not especially significant and is dropped. A second simplification is to carry all density in terms of initial density. A third simplification is to carry initial velocity head in the term: ( Ki). Application to Eqn. 2 of this method yields Eqn 4: (P)2 = (P)3 + ( Ki) V22/64.4 /144 (P) 2 = (P)3 + ( Ki) (w/(A ))22/64.4 /144 EQN. 4 application of Eqn 3a, V=(w/(A )) gives

=> (P) 2 = 2(P)3 + ( Ki) (w/(A))22/64.4 /144 & use density:

(PMP/(zRT)) 2 = (PM/(zRT))2(P)3 + ( Ki) (w/(A))22/64.4 /144 => (P2)2 -(P)2(P)3 -zRT( Ki) (w/(A))22/9274/M=0
Simplify the constants by using A in terms d4 to give: 1.273^2 *12^4/9274 = 3.62 (same as Crane Constant in 3.20-g)

P22 -P2P3 -3.6ZRT( Ki)(w)2/(Md4))2 =0 Ki)(w2)/(Y2Md4))=0

compare to Crane 3.20-g:

P22 -P2P3 -3.6ZRT(

The adiabatic expansion factor corrects for the simplifications made in derivation, as endorsed by APIRP521. The Crane method is given in example 3. The following five (5) validations are offered to confirm use of 36 Flare Nozzle for the listed pressure available and other requirements:

[12]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #1

RESULT OF FARRIS RV PROGRAM:

The Farris Relief valve program calculates a capacity of 4.22 MM#/hr for a 33 inch id tube at 25.9 psia pressure. Also provided are the equations used by the Farris Program to calculate capacity of flow nozzle for given conditions. The factor of 2 is because tube rupture flows from 2 ruptured sides. Capacity for 1 ruptured 33 inch tube is 2.06MM#/hr or 572#/sec. For these conditions density is: #/CF = 25.9*28.66/10.7/520 = 0.134 The area in square feet is: A, sf = (33/12)^2/1.273 = 5.94 For required mass flow of 2.06MM #/hr (572#/sec) velocity in feet/sec is: V= (#/sec)(cf/#)/(A, SF) = 572/0.134/5.94 V, fps = 719, & Mach# =719/1025= 0.70 This relief valve routine uses a method for flow at subsonic conditions thru a flow nozzle. The Farris RV Tube Rupture routine shows a 36 flare tip with 33 port is entirely adequate. It also shows that for required capacity, velocity need not approach sonic conditions.

[13]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #2

RESULTS FROM Daniel Orifice PROGRAM:

The Daniels Orifice program calculates a capacity of 2.102 MM#/hr for a 33 inch hole with 28 psia upstream pressure and 15psia downstream pressure. Daniels does not provide sizing equations used by the Daniels Program to calculate capacity of flow thru an opening at given conditions. For these conditions density is: #/CF = 26.4*0.9883*29/10.7/520/0.99 = 0.137 The area in square feet is: A, sf = (33/12)^2/1.273 = 5.94, square feet, SF For required mass flow of 2.05MM #/hr (570#/sec) velocity in feet/sec is: V= (#/sec)(cf/#)/(A, SF) = 570/0.137/5.94 V, fps = 700 Sonic Velocity is calculated by equation 5.7, Vs, fps = 44.8Root(T, R) =44.8520 =44.8*22.8 =1022 DANIELS Mach# at required capacity =V/Vs =700/1022= 0.68 An orifice bore of 33 inch at 26.4psia Matches Flare Criteria of Mach number less than 0.80 Pressure meets or exceeds requirement of critical pressure ratio Pressure drop is sufficient to meet flow requirements thru a 60 inch flare header.

[14]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #3 RESULTS FROM Crane TP410 Example 4.21: A comparison is made between results of TP-410, method of example 4.21 and results obtained by Daniels Orifice calculation method.
Determine inlet pressure to Nozzel discharging to atmosphere Crane (TP-410 25 th print 1991) formula for pressure loss in gas flow is given on page 3.4 by equation 3-20-g w =0.525Yd2(P/K)) 0.5 symbol w Y d P Pin Pout K 0.525 explanation lb/sec of flow expansion factor, pg A-22 id of pipe inches density in, #/cubic feet pressure drop psi P in-Pout psia psia
K resistance coefficient of system

Ex4.21 9.19 0.637 3.068 0.265 91.8 139.7 47.9 2.87 0.525

F3-36" 678.00 0.665 33.000 0.136 11.7 26.4 14.7 0.50 0.525

Dan Orif 678.00 0.665 33.000 0.136 11.7 26.4 14.7 0.50 0.525

dimensional term

=PM/(10.7ZR) P Pressure in Psia M molecular Weight Z Compression factor, ~1 Ro Temperature, 460+F k Cp/Cv by arrangement: 3.63Kw2/(d4Y2) = ( Pin-Pout) by density equation 3.63Kw2/(d4Y2) = MPin ( Pin-Pout)/(10.7Ro) (10.7Ro/M)3.63Kw2/(d4Y2) = Pin ( Pin-Pout) arranging to quadratic form gives

139.7 12.18 1.00 600.0 1.40

26.4 28.66 1.00 520.0 1.25

26.4 28.66 1.00 520.0 1.25

Pin ( Pin) - ( Pin)(Pout) - (10.7Ro/M)3.63Kw2/(d4Y2) =0 This equation can be solved by xcel solver for Pin given Pout Spreadsheet solution validated below for example 4.21, pg 4-14 Pin 139.7 26.4 26.4 Pout 47.4 14.7 14.7 #/sec 9.2 678 678 MW 12.2 28.7 28.7 T,R 600 520 520 id in f(P) den V fps M dP/P1 3.07 0.0 0.265 675.3 0.37 0.66 TP410-E4.21 33.00 0.0 0.136 839.3 0.79 0.44 Tan Flare F3 33.00 0.0 0.136 839.3 0.79 0.44 Dan Orif

[15]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #4 Result from Bernoulli Energy Balance


Provided to right is flow solution based on Bernoulli energy balance method. The details of the method are developed in section one of this discussion. The Bernoulli energy balance is valid at any sonic condition, either sonic, subsonic, or supersonic. However the pressure change across tip will vary depending upon sonic condition. The Critical Pressure ratio is used to determine flow conditions, sonic, subsonic or supersonic. Determination of critical pressure is given by Equation 5.1 of Appendix 1. The density change of gas is based on the adiabatic method. The friction K is based on an expansion K of 1.0 and a contraction K of 0.40. The sum of these two K factors determines friction head based on upstream velocity.
Nozzel Bern'li Eqn. Flow in Flow out variable dimension value up value down d nozzel inch 33.00 use adiabatic D Barrel inch 36.00 den change viscosity'cp 0.01 0.014 gas vol Q mscfd 6.45E+05 6.45E+05 gas sg gas sg 0.99 0.988 mass flow #/sec 564.67 564.67 k Cp/Cv 1.25 1.250 Pup psia 26.45 Pdn psia 14.700 Temperature R 520.00 462.36 Pmax up psia 26.49 1/k Den adab =u(Pu/Pd) 0.136 0.0852 Area in Sq ft 5.94 Beta, Beta 4.00 D2 (beta, ) inch 132.000 A d/s out sq ft 95.05 Velocity fps 697.66 69.761 Frict K Ke=1 Kc=0.4 1.40 2 feet, V /2g Vhead 7557.89 75.57 P hea d feet, P/ 27955.90 24857.17 frict head ft. K*Vin2/2g 10581.04 adjust mass | sum = 35513.78 35513.78 till =0 =diff |= 0.00 en'gy balance
bernoulli basis (P/+V2 /2g)1 = (P/+V2/2g)2 + hf

The mass rate is adjusted until the sum of energy terms are equal, or difference of terms equals zero. As is shown by sum row to right

Mach Number by API method is as follows: Where D is tip diameter in feet, and P2 is tip pressure in psia. Other units need be used as imperial units for the constant to apply. The Mach for Bernoulli are
1.72E-5*(565*3600)/(26.5/(33/12)
2

))*(1*520/1.25/28.7)0.5

= 0.658

[16]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #5 Result from Choke Valve


Mr. Choke Valve is called since some think any pressure above atmospheric causes or implies Mach number greater than one, M>1. Mach number is a ratio between sound velocity in gas divided into gas flow velocity. Determination of Critical Pressure Drop or Critical Pressure Ratio, CPR

The CPR is the minimum pressure ratio for sonic flow, and for any larger pressure ratio, flow inside a nozzle is subsonic, M<1. The pressure ratio is conventionally defined as pressure upstream divided into downstream pressure. Since flow is always in a direction of decreasing pressure, pressure upstream will be larger than pressure downstream. At pressure ratio of 1, flow is zero. From thermodynamics CPR is determined by formula 5.1. d/D = 33/36 = 0.92

C = 0.92 + 0.33 + 0.09 = 1.34 Since Nre>1E6, use C =1.2, as per example

Nre = 20 (651,430)*0.9883/(0.01245*33) =3.2E7

[17]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #5 Result from Choke Valve, ctd


For Practical Example consider a flared gas with k=1.25 & as given to right the critical pressure ratio is 0.555. Since gas expands to pressure of 1 atmosphere, which at sea level is about 14.7 psia, the max upstream pressure is 14.7/0.555 = 26.5 psia Since pressure available at HP flare KO drum is 37 psia, say by time gas reaches flare nozzle pressure there is only 26 psia available; check subsonic: 14.7/26 = 0.57 > 0.555 subsonic exist Calculate Sonic velocity for 60F hydrocarbon gas by formula 5.7 above: =44.8(520) =1022fps Flare group limits tip V to 0.8M, & allowable V = 0.8*1022 = 817fps. Listed conditions: C=1.2, d=33 P1 =26 for listed Conditions Qsc = 683645>651430 of Table 1 k=1.25, Pd/Pu=0.57 SG=0.988 T=520 Calc. #/sec= (scf/d)(1mol/379scf)(MW,#/mol)(1d/86400sec)
Mr. Choke #/sec =683,645,000/379*28.66/86400 =598#/sec

For 33 inch nozzle the area is 5.94 square feet. Thus the actual cubic feet per second equals: Gas density at Nozzle Condition is lb/cf = PM/10.7/R = 26*28.66/10.7/520 = 0.134#/actual cubic feet Thus 33 inch diameter flare tip is sufficient to pass 650#/sec of 26.66MW gas at pressure of 26psia and temperature of 60F (520 Rankin) at Mach 0.80. Mr.Choke, what is your Mach#?, It is =
=0.8*598/650=0.74 to meet or exceed requirements of Maximum Upset case

ACFS =V*A =5.94*817 =4853 actual cubic of gas feet/ second #/sec =ACF/sec *#/ACF = 4853*0.134 = 650#/sec The requirements per Zink Table, below, are: 650E6 SCF/Day/(379SCF/MOL)*(28.66#/mol)/(1440min/day)/(60sec /min) = 569#/sec Mr. Choke Valve is free to leave the witness stand. Zink Cases

A summary of Mr. Choke Valve is that: The flare tip of F3 is actually 36 inches, but using 33 inch nozzle is sufficient to pass 598#/sec at M=0.74, the requirements are only 569#/sec, therefore F3 tip is sufficient for the specified duty. The Mach of Mr. Choke meets requirements from both capacity and Mach requirements.

[18]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Witness #6

Result From ISA Control Valve

Control Valve Sizing program may also be used with Cv charts to estimate capacity of a flare tip.

The above output is from Fisher sizing program, FSP14. The upstream pressure was selected to not exceed Mach 0.8. At the given pressure, flow, temperature, gas heat capacity ratio, gas specific gravity and mass rate, the program calculates a Cv value. Various manufacturer offer Cv charts. Shown at right is Cv chart for large full port gate valves. FSP14 calculated a required Cv of 44069. As seen from the Cv chart above a 24 gate valve port will meet the required Cv. As size increases, the Cv increases. The listed flare orifice size by Zink is 33 inch, so a 33 inch port is sufficient to pass 2,050,000#/hr or 570#/sec. Mach number may be estimate as dP/(dP choked) = 8.1/11.1 = 0.73. Pressure drop, dP is determined as inlet, 22.8 minus outlet, 14.7 = 22.8-14.7 = 8.1psid. Calculation by API formula with 33 inch nozzle & given conditions yields a Mach Number of 0.78. In summary, Mr. ISA C. VALVE, also has testified that a 33 inch flare nozzle is adequate to process the required flow. More details on calculation equations for subsonic valves are available from the Instrument Society of America, publication: S75.01-1985 (R 1995), Flow Equations for Sizing Control Valves
[19]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer Additionally, the following important considerations deserve merit: Pressure Drop 60 Flare Header, HP Max Upset Condition
The following hydraulic analysis show that for a 60 flare header with a run of 1244 feet of feed pipe from Flare K/O drum, a flare tip pressure of 31.22 psia is achieved with given flow conditions and starting at a listed pressure of 37psia at flare K/O drum. While only between 30.1 and 28 psia is required to properly flow the 36 inch nozzle.
L i n e I d e n ti fi c a ti o n Ca s e Pha se Fluid in line F l o w r a te V i s c o s i ty D en s i ty Tem p er a tu r e I n l et P r es s u r e S p ec fi c G r a v i ty (S G ) M ol W t C p /C v o r K M ACH Num be r R o u g h n es s F a c to r P i p e S i ze / S c h ed u l e I n s i d e D i a m eter V el o c i ty S o n i c V el o c i ty P re s s u re D ro p R e y n o ld 's N u m b e r 0 .0 1 2 6 0 .0 0 59 157 1041 0.124 14,566,132 Q ty F eet o f P i p e 9 0 d eg El l s 4 5 d eg El l s s o ft T's h a r d T's R ed u c er en tr a n c e l o s s ex i t l o s s To ta l Eq u i v Len g th Eq ft d P I n l et P r es s u r e I n l et El ev a ti o n O u tl et El ev a ti o n S ta ti c H ea d Li n e Lo s s es C o n tr o l V a l v e Eq u i p m en t To ta l p r es s u r e l o s s O u tl et P r es s u r e N o tes : O r 651.43 M M S C F D 3 1 .2 2 782 0 .9 7 2 2 .3 0 0 0 .0 0 0 .9 7 0 0 0 .9 7 2 1 .3 3 4 0 0 1 0 302 260 0 0 220 0 0 0 Ft Ft Ft Ft Ft Ft Ft Ft Ft psi psi ft ft psi psi psi psi psi psi 1460 1 .8 6 2 1 .3 3 0 0 0 .0 0 1 .8 6 0 0 1 .8 6 1 9 .4 7 7 1 0 0 1 XS inch ft/s ec ft/s ec p s i /1 0 0 ft 0 .0 1 2 6 0 .0 0 59 161 1041 0.127 14,566,132 Q ty 942 455 35 0 0 28 0 0 Ft Ft Ft Ft Ft Ft Ft Ft Ft psi psi ft ft psi psi psi psi psi psi 3 8 3 .2 2 .8 4 1 9 .4 7 0 0 0 .0 0 2 .8 4 0 0 2 .8 4 1 6 .6 3 4 0 0 0 0 s td inch ft/s ec ft/s ec p s i /1 0 0 ft 0 .0 1 2 4 2 .0 0 42 337 1041 0.742 20,556,179 Q ty 7 5 .2 308 0 0 0 0 0 0 Ft Ft Ft Ft Ft Ft Ft Ft Ft psi psi ft ft psi psi psi psi psi psi 6 0 .1 0 1 6 .6 3 0 0 0 .0 0 0 .1 0 0 0 0 .1 0 1 6 .5 2 0 0 0 0 0 s td inch ft/s ec ft/s ec p s i /1 0 0 ft U n i ts 6 0 " s ec ti o n Ga s HP F LA RE 2 7 ,1 4 3 ,0 0 0 0 .0 1 5 0 .1 9 1 60 2 2 .3 0 .9 8 8 2 8 .6 5 1 .2 GAS SCFH cP l b /ft3 d eg F psig 60" run Ga s HP F LA RE GAS SCFH cP l b /ft3 d eg F psig U n i ts U n i ts fl a r e p =b a r r el eq u i l Ga s HP F LA RE 2 7 ,2 6 8 ,0 0 0 0 .0 1 5 0 .1 7 6 60 1 9 .4 7 0 .9 8 8 2 8 .6 5 1 .2 GAS SCFH cP l b /ft3 d eg F psig U n i ts 3 6 fl a r e ti p Ga s HP F LA RE 2 7 ,2 6 8 ,0 0 0 0 .0 1 5 0 .1 6 1 60 1 6 .6 3 0 .9 8 8 2 8 .6 5 1 .2 0.604 0 .0 1 2 3 6 .0 0 36 500 1041 1.749 23,982,209 Q ty 6 0 0 0 0 0 0 0 Ft Ft Ft Ft Ft Ft Ft Ft Ft psi psig ft ft psi psi psi psi psi psig s td inch ft/s ec ft/s ec p s i /1 0 0 ft GAS SCFH cP l b /ft3 d eg F psig

2 7 ,1 4 3 ,0 0 0 0 .0 1 5 0 .1 8 6 60 2 1 .3 3 0 .9 8 8 2 8 .6 5 1 .2

Psia at nozzle
First; 1300 foot run is only necessary at design of 500BTU/HR/SF which is excessive for maintenance work and two, the 651mmscfd of gas flow for a plant producing only 50mmscfd is excessive. These concerns are detailed below by depressurization study and review of heat flux.

[20]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer Corrosion Issues: Exterior Corrosion, Just one Location

Corrosion Issues: [21]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Just a poor weld job, not result of corrosion. A more detailed inspection appears in order. Corrosion alone does not condemn equipment. Repair of corrosion is a major function of every T&I.

[22]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer Depressurization Evaluation Determination of Gas Volumes Used for Depressurization Evaluation

[23]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer Depressurization Evaluation
HP HP HP HP HP HP HP HP
D203 D204 D106 B D207 D209 D212 D213 D2

13.0 11.0 11.5 11.0 11.5 5.0 5.0 14.0

24.0 33.0 36.0 23.0 36.0 8.5 13.0 55.5

60.0 LP suct 3186.2 60.0 LP Disch 3136.7 665.0 confy 3740.0 145.0 P/L suct 2186.2 625.0 cont fd 2" 3740.0 650.0 bst suc 166.9 650.0 bst dsc 255.3 50.0 LPTs 1/2 8545.2

2 6372.3 2 6273.4 2 7480.0 2 4372.3 2 7480.0 2 333.9 2 510.6 2 17090.3

0.260 0.260 3.494 0.649 3.284 3.415 3.815 0.217

30 24 4 24 20 8 12 16

150 150 150 150 150 150 150 150

300 1376 300 866 300 15 300 866 300 591 300 80 300 198 300 368

0.432 5098 0.276 5019 0.004 5984 0.396 3498 0.158 5984 0.480 267 0.776 408 0.043 13672

2549 2509 2992 1749 2992 134 204 6836

26.0 25.6 409.9 44.5 385.3 24" 12"V 17.9 30.6 58.1

50 50 615 115 565 600 600

60 60 665 145 625 650 650 50

25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5

60 26.0 0.98 60 25.6 0.98 665 409.9 0.81 145 44.5 0.95 625 385.3 0.81 650 17.9 0.81 650 30.6 0.73 50 58.1 0.98

Summary of Volumes used for simulation


P range Pmx 70 175 650 1230 mol 259.7 272.6 1817.5 76.4 2426.2 7.78 #/cf mx 0.33 0.84 3.68 # max 6622 6951 46347 1948 250 CF, Vo 20269 8242 12602 z 0.982 0.951 0.810 0.725

In summary, there are only a fixed amount of vessels inside Tanajib Plant. At any given time these fixed number of vessels contain a finite amount of gas molecules.

50 to 70 150 to175 570 to 650 1000 -1230 sum all

These molecules are released during de-pressure operations. If the molecules are released at a high rate, then the time to de-pressure is reduced. If the molecules are released at a moderate rate, as set by release rate requirements of OIMs and SAES-B-058, then the rate of release is reduced. CLEARLY BOTH HIGH RELEASE RATES AND LONG RELEASE TIMES CANNOT BOTH BE SUSTAINED. Either 1)the release time is reduced with increase in release rate or 2)release rate is reduced with longer release time. This concept is explored in more detail under section titled, Heat Release Rate. Heat exposure is a time sensitive factor and it is not possible to consider heat exposure flux without making consideration of the time of heat exposure. SAES B-058 7.4 Process vessels shall be designed with systems to de-pressure to 50% of the vessel's design gauge pressure within 15 minutes if:... and..b) The vessel is designed for pressures equal to or greater than a gauge pressure of 1725 kPa (250 psig). The depressurization sequences from the OIM of G54 are detailed below. Page 29 of G54 OIM 34.085 1. at time t=0 o Atmospheric Compressor K-101 HP Flare Valve NV-675 opens o LPPT Compressor K-102 HP Flare Valve NV-747 opens o Pipeline Compressor K-103 HP Flare Valves NV-809A and NV-809B open 2. At time 60 seconds o Atmospheric Compressor K-201 HP Flare Valve NV-710 opens, if train 2 is operational o LPPT Compressor K-202 HP Flare Valve NV-777 opens, if train 2 is operational o Pipeline Compressor K-203 HP Flare Valves NV-846A and NV-846B open, if train 2 is if operational 3. At time 250 seconds o Booster Compressor K-105 HP Flare Valves NV-1097 and NV-1050 open o Booster Compressor K-205 HP Flare Valves NV-1098 and NV-1079 open if train 2 is operational.

[24]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer Internal Depressurization Simulation results MAX RATE CASE to meet SAES B-058 7.4

The output of this simulation is shows a maximum Mach number of 0.45 is

[25]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer Depressurization Evaluation Dynamic Simulation of Depressurization using RO gas flow Equation. Plant G80 + R54 This depressurization Model Indicates Maximum Flow at flare is only 250#/sec, (900,000#/hr) not the 2 million pound per hour claimed on PFD. This is conservative model as Pressures on PFD at not achieved under current operation. Basis as per above Table and condensed to 4 pressure ranges as follows: P range 50 to 70 150 to170 570 to 625
1000 to1450

mol 135.7 407.9 738.4 64.2 1346.2

Pmx 70 170 625 1450

#/cf mx 0.36 0.88 3.24 7.51

# max 4344 13053 23628 2056

CF, Vo 11989 14833 7304 274

sum

Based on this evaluation Tanajib HP Flare emergency relief case is over specified by a factor of 100%, if requirements of SAES B-058 7.4. are applied.
In conclusion a 36 flare nozzle was calculated to run at 0.82 Mach by P&CSD based on the relieving rate of 2.05 MM#/HR to meet the requirement of 0.80 Mach number the flow need only be relaxed to 2.00 MM#/HR. The 2.05MM#/HR was based on specification of Pressure requirements for Operating conditions which are no longer valid.. This over specification applies to both Conditions shown on PFD & Current Operational Conditions.

[26]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

SAES_F-007 Table 1 on pg. 10 of 26

Clearly based on criteria listed in F007, Table 1, at heat exposure rates between 1500 BTU/HR/SF and 2000BTU/HR/SF an exposure time between one minute and several minute are permissible. The below simulation outputs indicate that if PFD maximum release rate of 570#/sec is used, the radiation exposure time between 1500 BTU/HR/SF and 2000 BTU/H/SF is about 60 second with Flare D of 770 feet...

[27]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

[28]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

[29]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

[30]

Considerations on Flare Parameters by Otis Armstrong, Licensed Professional Engineer

Summing up:
In summary this detailed review finds no reason a 36 Flare Tip cannot be used for this application of a temporary Flare.

Justifications are repeated as: 1. Less than 2% disagreement between standards requirements and P&CSD calculated values. P&CSD calculated Mach number is 0.82 as compared to the limit of 0.80. 2. Current Operational conditions do not require same relief capacity as shown on PFD because current operational pressures are lower than the pressures listed on PFD. 3. In house study indicates original sizing basis used more than 100% safety factor. 4. Internal Considerations for relieving Mach number indicate value less than 0.8 is to be expected.

[31]

You might also like