Adventures in Flaring

Paper from the AFRC 2015 conference titled Adventures in Flaring

2014 was a very exciting year for folks who operate flares. 2015 promises just as much excitement. There are new data from flare tests becoming available that help us understand flare performance over a range of operating conditions. EPA has issued the proposed Refinery Risk and Technology Review MACT revision rule. It certainly contains some interesting provisions. Finally, EPA also issued draft revisions for AP-42 emission factor that are, to say the least, surprising. This presentation will provide a brief overview and update of some of the main issues facing flare operators today. Specifically, it will answer the questions…; •; What is this hydrogen/olefin effect EPA is concerned about in the RSR Rule? Is it real?; ; •; What exactly do EPA's flare velocity limits protect us from? Why are they there? Is it time for them to go?; ; •; Will the proposed RSR Rule make it harder to avoid smoking flares?; ; •; Is your flare really about to emit 43 times more NOx than it did yesterday? The bizarre world of AP-42 flare emission factors.

Adventures in Flaring What EPA Didn't Tell You Presented by Scott Evans, Clean Air Engineering www.cleanair.com Hydrogen/Olefin Effect Velocity Limits Visible Emissions AP-42 Emission Factor Hydrogen and Olefins TCEQ Test - PFTIR Desktop Flares The Only Data on Hydrogen and Ethylene But what about the chart EPA produced in their supporting document? Doesn't this clearly show a hydrogen/olefin interaction? Petroleum Refinery Sector Rule: Operating Limits for Flares Dangling Blue Dots of Doom Figure 2 Evans - Clean Air Engineering Evans - Clean Air Engineering Hydrogen/Olefin Interaction Minute Data Plot H/O Interaction No H/O Interaction Combustion Zone Net Heating Value (Btu/scf) Combustion Efficiency (fraction) Combustion Efficiency (fraction) Minute Data Plot H/O Interaction No H/O Interaction Combustion Zone Net Heating Value (Btu/scf) Calculation Errors* Some Calculation Errors Corrected No Data Segregation High Propylene Data Removed Minute Data Not Representative* Remaining Data Points Are Noise * See Evans, S., Spellicy, R., "EPA PFTIR Data Analysis Critique", August 2014 Evans - Clean Air Engineering Evans - Clean Air Engineering Velocity Limits SVG = 0.17 SVG = 0.15 cf 57 /s st tu Te 4 B 29 cf 51 /s st tu Te 9 B 30 Only two data points for low-Btu steam-assisted flares McDaniel [83] Propylene/Nitrogen 60 ft/s , 30 0B tu/ scf Evans - Clean Air Engineering Pohl [84] 1027 Test 99 99.97% Destruction Efficiency 428.2 Evans - Clean Air Engineering McDaniel [83] Propylene/Nitrogen 60 ft/s , 30 0B tu/ scf Evans - Clean Air Engineering 4] gen 8 [ hl itro Po pane/N Pro McDaniel [83] Propylene/Nitrogen 60 ft/s , 30 0B tu/ scf Evans - Clean Air Engineering 40 0f t/s, 10 00 Btu /sc f 4] gen 8 [ hl itro Po pane/N Pro McDaniel [83] Propylene/Nitrogen 60 ft/s , 30 0B tu/ scf Evans - Clean Air Engineering EPA promulgates flare requirements in 40 CFR 60.18 51 FR 2701 January 21, 1986 Evans - Clean Air Engineering 40 0f t/s, 10 00 Btu /sc f 4] gen 8 [ hl itro Po pane/N Pro log10 (Vmax ) McDaniel [83] Propylene/Nitrogen 60 ft/s , 30 0B H T + 28.8 ) ( = 31.7 tu/ scf Evans - Clean Air Engineering 40 0f t/s, 10 00 Btu /sc f 4] gen 8 [ hl itro Po pane/N EPA [86 ] Pro McDaniel [83] Propylene/Nitrogen 60 ft/s , 30 0B log10 (Vmax ) H T + 28.8 ) ( = 31.7 tu/ scf Evans - Clean Air Engineering X X Known Stable Unknown Unstable ] 4 8 [ hl EPA [86 ] Po McDaniel [83] Evans - Clean Air Engineering CE >96.5 CE <96.5 Current EPA Velocity Limit High Velocity Flaring is a Good Thing Photo: Roozbeh Feiz Visible Emissions Visible Emissions Combustion Efficiency vs Visual Rating A smoking flare does not increase HAP emissions. Incipient Smoke Point In general, flares at or above the incipient smoke point are operating with high combustion efficiency. Evans - Clean Air Engineering Visible Emissions Analysis of flare soot conducted during the 2010 TCEQ flare study shows that the composition of flare soot is not comparable to diesel truck or generator soot. The organic carbon fraction is very small. E. C. Fortner, et al, Particulate Emissions Measured During the TCEQ Comprehensive Flare Emission Study, Industrial and Engineering Chemistry Research, 2012, 51 (39), pp 12586-12592 Evans - Clean Air Engineering Visible Emissions Current Language for M22 Section 3.5: Smoke emissions means a pollutant generated by combustion in a flare and occurring immediately downstream of the flame. Smoke occurring within the flame, but not downstream of the flame, is not considered a smoke emission. Suggested Revised Language for M22 Section 3.5: Smoke emissions means visible emissions persisting beyond one flame length from the visible flame tip. Smoke occurring within the visible flame is not considered a smoke emission. Evans - Clean Air Engineering Combustion Zone > Visible Flame All Soot Is Not Created Equal AP-42 Emission Factor Emission Factors Pollutant Old Factor New Factor Total Hydrocarbons 0.14 lb/mmBtu 0.14 lb/mmBtu Soot 0-274 µg.m3 0-274 µg/m3 Speciated Organics Various Limits Gone VOC None 0.55 lb/mmBtu Carbon Monoxide 0.37 lb/mmBtu 0.34 lb/mmBtu NOx 0.068 lb/mmBtu 2.9 lb/mmBtu Evans - Clean Air Engineering Emission Factors 3 Issues 1. EPA cites the PFTIR flare test reports but does not use the results from those tests. Instead they recalculate results from the raw data. EPA's recalculations are filled with errors and misinterpretation of the PFTIR data. The net effect is of these errors is to generally LOWER combustion efficiency for many if not most flares. 2. EPA assumes that NOx is solely a function of CO2 produced during combustion. Hydrogen, of course, also produces NOx when combusted. Therefore by EPA's method, NOx emissions approach infinity as H2 concentrations in the vent gas approach 100%. 3. The PFTIR raw data reports NO and NO2 values. However, the PFTIR was never calibrated for either compound. Therefore all the data EPA used to calculate the NOx emission factor are invalid. Evans - Clean Air Engineering Emission Factors For their NOx emission factor, EPA is… Performing incorrect calculations Using an improper model With invalid data Their incorrect calculations and assumptions regarding the PFTIR data will also affect other factors Evans - Clean Air Engineering Emission Factors NOx Emissions (lb/mmBtu) vs CE Usable data is available from the extractive tests performed during the 2010 TCEQ test. This data suggests the NOx factor should be LOWER than the current factor. 0.0096 + (-0.0266 / ((-100) + CE)) NOx = 0.0172 @ 96.5% CE NOx = 0.0229 @ 98% CE NOx = 0.0362 @ 99% CE vs NOx = 0.068 Current Limit Source: TCEQ 2010 Flare Test Extractive Data - propane steam flare tests (NOx/Total C basis) Evans - Clean Air Engineering Conclusion… A Much Simpler Solution… Questions? Scott Evans Clean Air Engineering [e-mail]