using isis-3d to simulate combustion performance in...
TRANSCRIPT
Using ISIS-3D to Simulate Combustion Performance in Flares
Ralph J. Price, PhDJoseph D. Smith, PhDAhti Suo-Anttila, PhD
Alion Science & TechnologyOwasso, OK
2009 ACERC Conference
Outline
• ISIS-3D Overview• Features• Models
• Applications• Flare Tips in Cross Wind• Flare Field Performance
• Conclusions
2009 ACERC Conference
ISIS-3D CFD Code
• Computational Fluid Dynamics (CFD) code with radiative heat transfer and combustion chemistry (developed by Alion)
• Developed to simulate heat transfer from pool fires and other buoyancy-driven flames
• Software capable of simulating complex, three-dimensional objects engulfed in fires
• Provides reasonably accurate estimates of total heat transfer to objects from large fires
• Predict general characteristics of temperature distribution in a flame and any nearby objects
• Accurately assess impact of variety of risk scenarios (wind, % flame coverage, thermal fatigue for given geometry, etc.)
• Reasonable CPU time requirements on “standard” desktop LINUX workstation
2009 ACERC Conference
ISIS-3D Limitations
• Sacrifice generality (large fires only) in favor of quick turnaround time and quantitative accuracy
• Reaction kinetics/radiation heat transfer models apply to large fires
• Reaction kinetics/radiation heat transfer models need to be calibrated using experimental data for a specific fuel
• Models and algorithms built to make ISIS-3D predictions reasonable when coarse grids employed
2009 ACERC Conference
Turbulence
• Turbulence is based on an eddy diffusivity concept• Most commonly used model is a very large eddy simulation
(VLES) model• ISIS-3D predicts puffing of fires and other large-scale
turbulent phenomena• Turbulence is mainly used to predict the eddy break-up time
used by the combustion model
2009 ACERC Conference
Radiation Inside Large Fires• High soot volume fractions make areas inside large fires non-
transparent (optically thick) which causes the flame to radiate as a cloud (radiatively diffuse)
• Fire volume defined as areas where soot volume fractions are greater than a specified minimum volume fraction (fSoot > fmin)
• Flame edge (fFlameEdge) defined where soot volume fraction is 0.05 ppm - based on comparisons with large fire experiments
Calculated flame surfaces for 3 time steps from ISIS-3D simulation of a pool fire
2009 ACERC Conference
• When fSoot < fFlameEdge = non-participating medium
• View factors from the fire to the un-engulfed surfaces of an object calculated at each time step
• Radiation view factor from object surface to surroundings calculated at each time step
• εFireSurface = 1 (fire is black body radiator)• Radiation from the fire surface to the
surroundings employed Tsurround = (constant ambient temperature)
Radiation Outside of Large Fires
2009 ACERC Conference
Combustion Model• Based on turbulent flame model of Said, et al. (1997) • Species simplified into 5 species
F = Fuel Vapor (from evaporation or flare tip) O2 = Oxygen (+ inert N2) PC = H2O(v) + CO2
C = Radiating Carbon Soot IS = Non-radiating Intermediate Species
• Combustion modeled using up to various reaction mechanisms• Model includes eddy dissipation effects and local equivalence
ratio effects• Reactions based on Arrhenius kinetics
• C and TA determined for all reactions
2009 ACERC Conference
Combustion Model Reactions
• PATH 1• Incomplete Fuel Combustion (soot producing)
• Fuel + O2 → Soot + PC + Heat• Soot Combustion
• Soot + O2 → CO2 + Heat
• PATH 2• Fuel Pyrolysis
• Fuel + O2 → Intermediate Species + PC + Heat• Combustion of Intermediate Species (soot producing)
• Intermediate Species + O2 → PC + Soot + Heat• Soot Combustion
• Soot + O2 → CO2 + Heat• Coefficients chosen so that complete combustion of C and IS produce same
species and thermal energy as direct complete combustion of fuel
2009 ACERC Conference
Combustion MechanismsPropane:
Ethylene:
Mixed Gas:
2009 ACERC Conference
Low Profile Flares - Modeling Issues• High tip velocity increases air
entrainment• Tip design critical to air entrainment• Local high velocity can translate into
high sound levels• Assist media not available to
increase combustion air• smoke below certain tip pressure (D-
stage pressure)• Tip spacing critical
• Flares must cross light• Possible Flame merge lengthens
flames• Adjacent rows compete for air (longer
flames, poor performance)
2009 ACERC Conference
Approach to Modeling Full Flare Fields• Model Single Burner Test
• Perform Calibration Tests• Calibrate Soot Yield and Kinetic Parameters for Test Fuel• Predict flame shape and size
• Model Multi-Burner Test• Perform Radiation Calibration Tests• Validate kinetics by comparing model to experimental data (radiation
data, flame shape, flame height)
• Model Full Flare Field• Use Calibrated Soot Yield and Radiation Models• Predict Flare Performance (Smoke Production/Air Demand)• Predict Radiation Load on Wind Fence
2009 ACERC Conference
Single Flare: Wind vs. No-Wind
No Wind Condition: Long “pencil-like” flame
Wind
y Con
dition
s
8-10 mph Wind Condition:Tilted shorter bushy flame
Propane Flare
2009 ACERC Conference
Single-burner MeshRectangular cellsLocal refinement near burner tip110,000 computational cells
2009 ACERC Conference
Flame in No Wind (Propane)
No wind produces tight “pencil-like” flame
13 m
16 m
2009 ACERC Conference
Flame in No Wind (2D Slice)
2009 ACERC Conference
Flame in 3.0 m/s Wind
Wind produces tilted bushy, shortened flame
Side View
2009 ACERC Conference
15 m (49 ft)
20 m (66 ft)
Flare tip located 2 m (~7 ft) above grade
0.05 ppm Soot Iso-surface used to represent Flame volume – surface colored by gas temperature
Predicted Flame Length for No-Wind
2009 ACERC Conference
Single-Tip Ground Flare Test ResultsMeasured flame length wo/ wind
13.8 -15.3 m (48’-53’)
• No Wind Condition (<1 mph wind)• Ave Flame Length = 14.8 - 16.3 m
(48 - 53 ft)• “Pencil-like” tight flame• Small non-luminous flame at base• Propane Flow rate: measured 1.4”
WC @ 57 ˚F across orifice plate => 7.3 psig tip back pressure (measured on 18 inch pipe run)
Non-luminous Flame ~1 m
2009 ACERC Conference
10.5 m (34 ft)
~8-10˚ flame tilt
Single-Tip Ground Flare Test ResultsMeasured flame length w/ wind • 1.4” WC @ 57 ˚F => 7.3 psig
tip pressure
• 12-16 km/hr (8-10 mph) crosswind
• ~30% flame height reduction
• Minimum flame tilt (~8º)
2009 ACERC Conference
Model Results
• Air demand based on total mass flow through a 3.6 m square plane located 20 m height above flare
• Predicted 60 kg/sec air demand by flame• Predicted flame height is 17 m above ground (15 m flame
length)• Radiation depends on soot, CO2, and H2O concentrations in
flame and flame size• Soot yield from hydrocarbon assumed constant for propane• Predicted approximately 3 MW radiation loss from 22 MW
Flame or 13.6% heat loss
2009 ACERC Conference
Three-burner MeshRectangular cellsDomain size is 30 m X 35 m X 25 mLocal refinement near burner tips and radiation meters188,000 computational cells
2009 ACERC Conference
Predicted 3-Burner Flare (Ethylene)
2009 ACERC Conference
~16 m (52 ft)
Predicted Flame Height for 3-Burner Test
12 m (39 ft)
2009 ACERC Conference
Predicted Flame Height for 3-Burner Test
2009 ACERC Conference
3-burner Flare Test with Ethylene (no wind)
Tip Height~3 m (10’)
Non-luminous region ~1 m (3’)
Flame height ~11 m (~36’)
2009 ACERC Conference
Multi-Tip Ground Flare
Test Problem• 124 TPH Propane• Wind Fence surrounds ground flare• Specified open area in fence• Individual Flare Tips located above grade • No wind conditions• Low pressure/non-momentum driven flames
Flames merge due to side air demand from
sides of rows
2009 ACERC Conference
Full Field Flare Grid
Domain size is 10 m beyond wind fence and 25 m highLocal refinement near burner rows/tips700,000 cells (Sustained Flow)1,200,000 cells (Peak Flow)
2009 ACERC Conference
Multi-Tip Ground Flare Results (Propane)
2009 ACERC Conference
Peak Flow - No Wind Condition Results (Mixed Gas)
Predicted flames well below total fence height
2009 ACERC Conference
Fuel Concentrations for Peak Flow (Mixed Gas)
Side view along row showing flame elongation toward center of row
End view of flame height for each row and impact of inflow on outer rows
2009 ACERC Conference
Summary of all Cases for “no-wind” conditionsCase
Description Evaluation Plane
Area (m2) Total Air Flow
(kg/s) Fuel Flow
(kg/s) Air/Fuel
Mass Ratio 1 Burner Propane 14.63 60 0.46 130
1 Burner (Tip 3) Ethylene 13.26 52 0.94 55
3 Burner (Tip 3 - 7.3PSI)
Ethylene 36
134 2.88 47
3 Burner (Tip 3 - 11.4PSI)
Ethylene 36
144 3.84 38
3 Burner (Tip 4 - 7.3PSI)
Ethylene 36
150 4.26 35
3 Burner (Tip 4 - 11.4PSI)
Ethylene 36
160 5.79 28
Full Field Peak Flow Ethylene
3226 9700 262.3 37
Full Field Sustained Flow
Mixed Gas 1843 4800 93.6 51.3
2009 ACERC Conference
Wind
Application: Utility Flare in Purge Conditions
2009 ACERC Conference
Application – Flare Start-Up
2009 ACERC Conference
Application – Pool Fire Analysis
2009 ACERC Conference
Conclusions
• Modeled flame shape/size for 3 fuels for single and three burner tests
• Predictions compared to data (2 tip sizes, 3 operating pressures, 2 radiation sample locations)
• Predicted “reasonable” estimates of radiation heat transfer and air demand for low profile flare (single burner and 3 burners)> Air/fuel ratios range from 28 to 47 for 3-burner test and
from 37 (Peak Flow Case) to 51 (Sustained Flow Case)
• Calibrated flare model applied to full-flare field to estimate:> Air demand for specified tip/row spacing> Radiation load on wind fence for nominal and peak flow cases> Expected flame height and smoke production for nominal and peak flow cases
2009 ACERC Conference
Acknowledgements
• Zeeco – Flare Designer• Scot Smith – Zeeco• Ahti Suo-Anttila – Alion Science & Technology• Joseph Smith – Alion Science & Technology