Texas commission on environmental quality 2010 flare study

TEXAS COMMISSION ON ENVIRONMENTAL QUALITY Texas Commission on Environm ental Q uality 2 0 1 0 Flare Study Flare Task Force Presented by Air Quality Division Presented to American Flame Research Committee (AFRC) Combustion Symposium and Industrial Flares Colloquium, Houston, TX September 19, 2011 Overview • • • • • Objectives and Limitations Study Design Results Additional Flare Research Results Summary TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 2 Objectives Assess the impact of high turndown (low flow) rate of vent gas on flare destruction and removal efficiency (DRE) and combustion efficiency (CE) - DRE is the percent removal of hydrocarbon from flare vent gas. - CE is the percent of hydrocarbon in vent gas converted to carbon dioxide. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 3 Objectives • Assess if flares operating within 40 Code of Federal Regulations (CFR) §60.18 achieve the assumed hydrocarbon DRE of at least 98% at high turndown, varying assist ratios, and vent gas heat content • Identify and quantify the hydrocarbon species in flare plumes TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 4 Limitations • Limited vent gas composition: Tulsa natural gas, propylene, and nitrogen - Propane was used for limited test runs. - Hydrogen was not included in any test run. • Two flare tip sizes and assist configurations were tested. • High turndown (low flow) operating conditions were focus of study. • Study was not designed to evaluate: - Flare operations under upset or emergency conditions - Hydrogen flares - Flares specifically designed for routine, low flow applications TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 5 Study Design: Flare Tips • 36-inch steam-assisted flare with upper and center steam assist - Upper steam carries ambient oxygen into the combustion zone to prevent smoke. - Center steam helps push the combustion outside of the tip. • 24-inch air-assisted flare The fan motor had a variable frequency drive capable of small air flow adjustments by the control room. • Both flare tip designs are commonly used for routine low-flow vent gas streams TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 6 • Flare configurations tested represent flares commonly used in both routine process and emergency service (dual service). • For these dual-service flares, 2009 TCEQ emissions inventory data indicates: - 21% of dual-service flares are air-assisted. Of these air-assisted, dual-service flares: 77% are 12 to 36 inches in diameter and represent 95% of total 2009 emissions for dual-service, airassisted flares. - 45% of dual-service flares are steam-assisted. Of these steam-assisted, dual-service flares: 41% are 24 to 48 inches in diameter and represent 57% of total 2009 emissions for dual-service, steam-assisted flares. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 7 Representativeness • All test points are within 40 CFR §60.18 criteria. • Tip velocities are similar to those observed during field testing at Marathon facilities. • Steam-assist rates are representative of manufacturer recommendations for this size flare tip configuration. • Testing performed on stable flames, per John Zink staff. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 8 Operating Conditions • Vent gas streams with heat content of 350, 600, and 2,149 British thermal units per standard cubic foot (Btu/scf) 40 CFR §60.18 minimum heating value for an assisted flare is 300 Btu/scf. • Vent gas streams with low flow rate - 0.1% and 0.25% of rated design capacity - Steam-assisted flare = 937 lb/hr and 2,342 lb/hr - Air-assisted flare = 359 lb/hr and 937 lb/hr TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 9 Operating Conditions • Assist rates varied between zero assist to over assist near flameout (snuff point). • Measurements were taken at points between the incipient smoke point and near snuff point. Four to six points per test series with up to three repetitions per point • Tip velocity of vent gas, including center steam, was between 0.6 and 2.0 feet per second (fps). TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 10 Data Collection • Extractive measurements - Aerodyne Research: quantum cascade laser, proton transfer reaction mass spectrometer, gas chromatograph (GC), aerosol mass spectrometer, particle analyzers - TRC: GC • Remote sensing measurements - Telops: Field portable radiometric spectrometer - Industrial Monitor and Control Corporation: passive and active Fourier transform infrared (PFTIR and AFTIR) detectors - All rem ote sensing companies perform ed single­ blind m easurem ents. • Leak Surveys, Inc: FLIR GasFindIR infrared cameras TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 11 Extractive Sam pler • Forced air device designed by Aerodyne, the University of Texas at Austin Center for Environmental and Energy Resources (UTCEER), and John Zink • Extensive calibration procedures • Positioning of the sampler during measurements - Two FLIR GasFindIR infrared cameras - Temperature of three thermocouples - Visual line of sight by crew holding the position chains and project personnel in the control room - Oxygen and carbon dioxide measurements - Global positioning system coordinates TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 12 Extractive Sam pler TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 13 Extractive Sam pler In le t TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 14 Extractive Sam pler Flue gas eductor Extractive sample inlet Pitot Elevation chain Sample lines Positioning chains TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 15 TCEQ Extractive Sam pler Positioning TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 16 Extractive Sam pler During morning start-up procedures TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 17 Elevated Position TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 18 Test Point S4.4 TCEQ Vent Gas Btu/scf 2,342 lb/ hr 350 Btu Upper Steam Center Steam 327 lb/hr zero DRE Active FTIR mirror Industrial Monitor and Control Corporation (515' U l-8189 TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 19 Test Point A 4.6 Vent Gas Btu/scf DRE (%) 937 lb/hr 350 Btu 99.4 TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 20 Control Room Real-Tim e Observations and Measurem ents Extractive data measurements and waste gas flow rate I TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 21 Results: High DRE Measured • The flares tested were able to achieve greater than 99% DRE and CE for vent gas streams with low heating value at low flow rate conditions. • For the conditions tested, the highest DRE and CE was achieved at or near the incipient smoke point. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 22 Steam-Assisted Flare DRE Constant V en t Gas Flow Rate of 2 ,3 4 2 lb /h r Incipient smoke points ♦ 2149 Btu/scf ■ 600 Btu/scf * 350 Btu/scf S te a m -to -V e n t Gas Ratio lb /lb TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 23 Results: Heating Value Critical • Vent gas with high heating value (2,149 Btu/ scf) has a wider operating range for steam to vent gas ratios. • Vent gas with lower heating values (350 and 600 Btu/scf) has a narrower operating range for steam to vent gas ratios, regardless of tested flow rate. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 24 Results: Heating Value Critical TCEQ DRE Versus S te a m -to -V e n t-G a s Ratio a t Varying V e n t Gas Heating Values 100 DRE - Propylene (% ) 98 ♦ ♦ 96 94 ♦ Series S3: 350 Btu @ 937 lb/hr" 92 ♦ Series S4 : 350 Btu @ 2,342 lb/hr 90 ■ Series S5: 600 Btu @ 937 lb/hr 88 ■ Series S6: 600 Btu @ 2,342 lb/hr ♦ ♦ 86 84 0.00 0.50 1.00 S te a m -to -V e n t-G a s Ratio 1.50 TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 25 Results: Center Steam Im pacts DRE • At minimum recommended center-steam assist rates, the steam-assisted flare was not able to achieve 99% DRE for the vent gas stream of 350 Btu/scf at 937 lb/hr flow rate. Manufacturer recommended a minimum center steam operating range of 300-500 lb/hr. • At 937 lb/hr flow rate, a steam-to-vent-gas ratio of less than 0.25 (with zero center steam) was required to achieve 99% DRE. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 26 Test Point S3.1 Btu/scf Upper Steam Center Steam "Transparent" flame occurred when low DRE was measured TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 27 Test Point S3.6 TCEQ Vent Gas Btu/scf 937 lb/hr 350 Btu Upper Steam zero Center Steam DRE zero Visible orange-yellow flame occurred when high DRE was measured TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 28 Im p a c t of Center Steam TCEQ DRE fo r Tests S7 to S l l a t 3 5 0 B tu /s c f w ith Constant Steam Assist and Varying V en t Gas Flow Rate 100 95 ->w ♦ 90 ♦ ♦ 85 a> c 80 QJ > a 75 70 CL Q □ Total Steam = 500 lbs/hr, 0 lbs/hr Center (S8) i Total Steam = 1000 lbs/hr, 0 lbs/hr Center (S9) o LU < Total Steam = 1050 lbs/hr, 500 lbs/hr Center (S7) 65 Total Steam = 835 lbs/hr, 0 lbs/hr Center (S10) 60 55 Total Steam = 830 lbs/hr, 240 lbs/hr Center ( S l l ) 50 500 1500 2500 V en t Gas Flow Rate ( l b / h r ) 3500 TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 29 Results: Operating Conditions Critical Steam-assisted flare DRE measured at 98% under limited operating conditions when vent gas stream had low heating value and low flow rates. - The DRE and CE decrease almost linearly as steam assist rate increases. - As DRE decreases, flame becomes more "transparent." TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 30 TCEQ Test Point S4.2 Steam to Btu/scf VG Ratio TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 31 Test Point S4.7 TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 32 TCEQ Test Point S4.3 Btu/scf TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 33 Steam to VG Ratio DRE (%) DRE Versus Steam Assist Rate TCEQ DRE for Tests S5 and S6 at 6 0 0 Btu/scf with Constant Vent Gas Flow Rate 100 * 90 80 ♦ Vent Gas = 937 lbs/hr (S5) a> c 70 QJ > a 60 □ Vent Gas = 2342 lbs/hr (S6) 0 i_ CL 1 LU CL 50 40 O 30 20 0 1000 2000 3000 4000 Total Steam Assist ( l b / h r ) (C e n te r Steam Constant at 5 0 0 l b / h r ) TCEQ • 2010 Flare Study • AQD • September 19, 2011 • 5000 Page 34 Results: Air Assist Rates Critical The air-assisted flare DRE measured greater than 97% when the excess air factor was less than 10. - - Excess air factor: The amount of air in excess of what is required to achieve theoretical stoichiometric combustion represented as a factor. Example: 15 pounds of air is required to burn 1 pound of propylene. I f the air assist rate is 150 pounds of air per pound of propylene, the excess air factor would be 10. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 35 DRE Versus Excess Air Excess Air Factor of 10 100 Incipient smoke points DRE - Propylene (%) 98 ♦s ▲ • 96 < A5: 8 0 lb /h r ■ ■ Hydrocarbon 94 A6: 131 lb /h r Hydrocarbon 92 * i A3: 2 0 0 lb /h r Hydrocarbon 90 i A4: 3 3 0 lb /h r 1 88 0 • --------------- ,--------- A-------- ,------------S,--------------- ,--------------- , 10 15 20 25 30 35 Excess Air Factor TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 36 40 Hydrocarbon rj toq Results: Comparison to 1983 Steam Assisted Tests • TCEQ 2010 test points at 2,149 Btu/scf are similar to EPA 1983 test points at 2,183 Btu/scf. • TCEQ 2010 test points at 350 and 600 Btu/scf are significantly different than EPA 1983 test points. During the EPA 1983 test at vent gas heating values below 600 Btu/scf, the steam assist was not used and the tip was unassisted. High CE would be expected in this configuration. • An 8-inch upper steam-assisted tip was used in EPA 1983 tests. • EPA 1983 testing occurred during calm wind conditions. TCEQ testing occurred during variable wind conditions. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 37 TCEQ 2010 and EPA 1983 SteamAssisted Flare Test Data TCEQ 100 % Combustion Efficiency a EPA 1983 Test 80 75 □ TCEQ 2010 Points at 2,149 Btu/ scf 70 65 ❖ 60 OTCEQ 2010 Points below 600 s Btu/scf 55 50 2 3 4 5 6 S te a m -to -v e n t Gas Ratio TCEQ • 2010 Flare Study • AQD • June 14, 2011 • Page 38 TCEQ Results: How Well Do Passive Techniques Agree with Direct Measurement? Single-blind CE measurements from the PFTIR were comparable to the Aerodyne extractive CE measurements at higher CE conditions. Below 87% CE, some instances of poor correlation were observed. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 39 CE ( % ) TCEQ PFTIR Versus Extractive Measurem ents Test Points Ranked by CE TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 40 Results: Combustion Zone Gas Net Heating Value The TCEQ 2010 combustion zone gas net heating value data curve above 200 Btu/scf is very similar to recent Marathon and other EPA consent decree passive FTIR measurements. - Marathon measurement performed on different flare tip configurations and different vent gas flow conditions. - Marathon, Texas City, test on 24-inch diameter flare with center, upper, and lower steam and tested under refinery base load conditions at a tip velocity 1 to 3 fps. - Marathon, Detroit, test on 16-inch diameter flare with center and upper steam and tested under refinery base load conditions at a velocity of 2 fps. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 41 H Combustion Zone Gas Net Heating Value TCEQ 3 CE Versus Combustion Zone Gas Net Heating Value Combustion Efficiency (%) 100 m U f 98 96 94 92 90 \J !r •• ■ " A A 88 86 84 82 80 78 76 74 72 70 ♦ Test Series S3 Test Series S4 Test Series S5 • Test Series S6 68 66 64 62 60 0 100 200 300 400 500 Combustion Zone Gas Net Heating Value(Btu/scf) TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 42 600 Marathon Petroleum Company Passive FTIR Test Results TCEQ M arathon Petroleum Company Detroit Refinery - CP Flare (2010) Texas City Refinery - Main Flare (2009) to 2 p o 100% 98% 96% 94% 92% 90% 88% 86% 84% 82% 80% 78% 76% 74% 72% 70% 68% U 66% ^ ^ ^ u j= ~ UZ u c .2 D etroit and Texas City Comparison A Series Runs (Base Load) CE Versus Combustion Zone Gas Net Heating Value Texas City Detroit 64% 62% 60% - i-------------1 -------------1------------ 1------------ 1 -------------1 -------------1 ------------ 1 -------------1 0 Source: Marathon Petroleum Company 50 100 150 200 250 300 350 400 Combustion Zone Gas Net Heating Value (B T U /s c f) TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 43 450 Communication Plan • TCEQ is developing a cooperative flare training program with stakeholders. - Targeted to industry and regulators - Goals: ■ Share information from the TCEQ 2010 Flare Study ■ Improve knowledge and efficiency of flare operations • TCEQ is developing a presentation for local community advisory panels. The main objective is to communicate that a luminous and visible flame generally indicates that a flare is combusting hydrocarbons effectively. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 44 TCEQ-Funded Ongoing Flare Research University of Texas at Austin (UT) Center for Energy and Environmental Resources Air Quality Research Program projects - UT project to use multivariate image analysis and Fluent computational fluid dynamic (CFD) modeling software - Lamar University project to replicate TCEQ 2010 Flare Study Project results using CFD modeling software Lamar Supplemental Environmental Program project Use CFD modeling software to identify speciated VOC emissions from flaring operations of various chemical processes TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 45 Other Ongoing Flare Research • EPA PFTIR Projects EPA consent decree requirement to use PFTIR testing on flares at refineries and chemical plants • John Zink Pressure Assist Flare Testing - High CE at all vent gas flows (98-100% ) - Potential alternative to an air assist flare under specific vent gas flow conditions. - Does not conform to current 40 CFR §60.18 requirements • International Flaring Consortium Test Test results have not been released. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 46 Results Sum m ary • Air- and steam-assisted flares can efficiently control low Btu vent gas at low flow rates under limited operating conditions. - The assist-to-vent gas flow operating range to achieve greater than 98% DRE was limited. - A slow rolling, bright orange flame near the incipient smoke point was observed when DRE was measured to be greater than 98%. • Controlling flare assist rates is critical to achieving high DRE. - The assist-to-vent gas ratio operating range increased with increased heating value of vent gas. - Increasing vent gas flow rate had a marginal effect on the assist-to-vent gas ratio operating range. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 47 Results Sum m ary • A flare can be operated under 40 CFR §60.18 criteria and not achieve 98% DRE. • Flares were easily over-assisted. - Air-assisted flare with an excess air factor greater than 10 measured less than assumed 98% DRE. - Steam-assisted flare combusting 350 and 600 Btu/scf waste gas required a steam -to-vent gas ratio of less than 1.1:1 to achieve 98% DRE. - Steam-assisted flare combusting 2,149 Btu/scf waste gas required a steam -to-vent gas ratio of less than 3.3:1 to achieve 98% DRE. - Type of steam assist (center versus upper) impacts flare DRE. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 48 Contact In fo rm atio n • Danielle Nesvacil, Emissions Assessment Section [e-mail], [phone] • Russell Nettles, Emissions Assessment Section [e-mail], [phone] • Study results are posted at: - TCEQ's Flare Task Force Stakeholder Group Web site http://w w w .tceq.texas.gov/airquality/stationary-rules/ flare_stakeholder.html - Sign up for e-mail updates through TCEQ's GovDelivery listserver. Select "SIP Hot Topics" under the "A ir Quality" heading to receive Flare Task Force updates. TCEQ • 2010 Flare Study • AQD • September 19, 2011 • Page 49