Paul D. Ronney
Dept. of Aerospace & Mechanical Engineering
University of Southern California
Los Angeles, CA 90089-1453 USA
QuickTime™ and a
decompressor
are needed to see this picture.
Institute of Mechanics,
Chinese Academy of Sciences
March 22, 2001
http://carambola.usc.edu
Supported by NASA-
OUTLINE
• Motivation
• Time scales
• Examples
• Premixed-gas flames
• Nonpremixed gas flames
• Condensed-phase combustion
• Summary
• Challenges for future work
• Perspective on space flight training
University of Southern California - Department of Aerospace & Mechanical Engineering
MOTIVATION
• Gravity influences combustion through
• Buoyant convection
• Sedimentation in multi-phase systems
• Many experimental & theoretical studies of µg combustion
• Applications
• Spacecraft fire safety
 Better understanding of combustion at earth gravity
University of Southern California - Department of Aerospace & Mechanical Engineering
TIME SCALES - PREMIXED-GAS FLAMES
• Chemical time scale
tchem ≈ d/SL ≈ (a/SL)/SL ≈ a/SL2
(a = thermal diffusivity, SL = laminar flame speed)
• Buoyant transport time scale
t ~ d/V; V ≈ (gd(Dr/r))1/2 ≈ (gd)1/2
(g = gravity, d = characteristic dimension)
Inviscid: tinv ≈ d/(gd)1/2 ≈ (d/g)1/2
Viscous: d ≈ n/V  tvis ≈ (n/g2)1/3 (n = viscosity)
• Conduction time scale
tcond ≈ Tf/(dT/dt) ≈ d2/16a
• Radiation time scale
trad ≈ Tf/(dT/dt) ≈ Tf/(L/rCp)
Optically thin radiation: L = 4sap(Tf4 – T∞4)
(ap = Planck mean absorption coefficient)
 trad ~ P/sap(Tf4 – T∞4) ~ P0, P = pressure
University of Southern California - Department of Aerospace & Mechanical Engineering
Time scales (hydrocarbon-air, 1 atm)
T
Tiim
mee ssccaallee
C
Ch
heem
miissttrryy ((ttcchheemm))
o
orr d
diiffffu
ussiio
on
n ((ttddiiffff))
B
Bu
uo
oyyaan
ntt,, iin
nvviisscciid
d ((ttiinnvv))
B
Bu
uo
oyyaan
ntt,, vviisscco
ou
uss ((ttvviiss))
C
Co
on
nd
du
uccttiio
on
n ((ttccoonndd)),, d
d == 55 ccm
m
R
Raad
diiaattiio
on
n ((ttrraadd))
S
Stto
oiicch
h.. ffllaam
mee
00..0000009944 sseecc
L
Liim
miitt ffllaam
mee
00..2255 sseecc
00..007711 sseecc
00..001122 sseecc
00..9955 sseecc
00..1133 sseecc
00..007711 sseecc
00..001100 sseecc
11..44 sseecc
00..4411 sseecc
• Conclusions
• Buoyancy unimportant for near-stoichiometric flames
(tinv & tvis >> tchem)
• Buoyancy strongly influences near-limit flames at 1g
(tinv & tvis < tchem)
• Radiation effects unimportant at 1g (tvis << trad; tinv << trad)
 Radiation effects dominate flames with low SL
(trad ≈ tchem), but only observable at µg
• Radiation > conduction only for d > 3 cm
• Re ~ Vd/n ~ (gd3/n2)1/2  turbulent flow at 1g for d > 10 cm
University of Southern California - Department of Aerospace & Mechanical Engineering
Premixed-gas flames – flammability limits
• Too lean or too rich mixtures won’t burn
- flammability limits
• No limits without losses – no purely chemical criterion
(Mukunda et al., 1990; Giovangigli & Smooke, 1992)
• Models of limits due to losses - most important prediction:
burning velocity at the limit (SL,lim)
• Heat loss to walls: tchem ~ tcond  SL,lim ≈ 40a/d (Pelim = 40)
(Spalding, 1957; Jarosinsky, 1983)
• Upward propagation: rise speed at limit ~ (gd)1/2 (Levy, 1965);
causes stretch extinction (Buckmaster & Mikolaitis, 1982):
tchem ~ tinv  SL,lim ~ (ga2/d)1/4
• Downward propagation – sinking layer of cooling gases near
wall outruns & “suffocates” flame (Jarosinsky et al., 1982)
tchem ~ tvis  SL,lim ≈ 1.3(ga)1/3
University of Southern California - Department of Aerospace & Mechanical Engineering
Premixed-gas flames – flammability limits
10
10
4
Le=0.17
Le=0.27
Le=0.71
Le=0.96
Le=1.47
Le=1.74
Le=1.88
3
10
Le=1.97
Le=2.19
Le=2.38
Le=3.22
Le=3.34
Le=3.57
Peclet number at limit
Peclet number at limit
10
2
Pe = 40
Pe = 1.6 5 Gr
10
1/3
10
10
10
2
10
3
10
4
5
6
10
10
Grashof number
10
7
10
8
10
Downward propagation
9
Le=0.17
Le=0.3
Le=0.71
Le=0.86
Le=0.96
Le=1.46
4
Le=1.73
Le=1.86
Le=2.02
Le=2.17
Le=2.76
Le=3.2
3
2
Pe = 40
1
10
5
10
Pe = 0.2 3 Gr
1
10
2
10
3
10
4
5
6
1/2
10
10
10
Grashof number
7
10
8
10
9
10
10
Upward propagation
Wang & Ronney, 1993
University of Southern California - Department of Aerospace & Mechanical Engineering
Flammability limits – losses - continued…
• Big tube, no gravity – what causes limits?
• Radiation heat loss (Joulin & Clavin, Buckmaster, 1976)
1.2L f
1
SL ,lim 
r Cp
Tf
(no reabsorption)
• Doesn’t radiative loss decrease for weaker mixtures, since
temperature is lower? NO!
Heat loss
T2
Impact of heat loss ~
~ -E/RT  as T 
Heat release
e
• Predicted SL,lim consistent with µg experiments (Ronney,
1988; Abbud-Madrid & Ronney, 1990)
University of Southern California - Department of Aerospace & Mechanical Engineering
Reabsorption effects on premixed flames
• Is radiation always a loss mechanism?
• Reabsorption may be important when aP-1 < d
-1
-1
atm )
• Blackbody particles - extend limits & increases SL (Eudier &
Joulin, 1988; Abbud-Madrid & Ronney, 1993)
• Gases – some radiation still escapes (Ju, Masuya, Ronney 1998)
» Absorption spectra of products different from reactants
» Spectra broader at high T than low T
» Dramatic difference in SL & limits compared to optically thin
University of Southern California - Department of Aerospace & Mechanical Engineering
Premixed-gas flames - stretched flames
• Nonuniform flow, unsteady/curved flames: “flame stretch”
1 dA
S
(A = flame area)
A dt
• Strong stretch (S-1 ≈ tchem) extinguishes flames
• Moderate stretch strengthens flames for Le < 1
Thermal diffusivity of the bulk mixture (
a)
Le 
Mass diffusivity of scarce reactant into the bulk mixture (D)
University of Southern California - Department of Aerospace & Mechanical Engineering
Premixed-gases - Self-Extinguishing Flames
• Spherical expanding flames, Le < 1: stretch allows flames
to exist in mixtures below radiative limit until rf too large &
curvature benefit too weak


1 dA
1 d
2 drf
dS
2S
2
2
2
S

4

r



S
ln
S

Q
f
2
A dt 4rf dt
r f dt
dR
R
80
Q = 0.36
Dimensionless radius (R)
70
60
Q = 0.4
50
40
Q = 0.42
30
Q = 0.45
20
Q = 0.5
10
0
0
20
40
60
Dimensionless time (
80
100
)
 Dual limit: radiation at large rf, curvature-induced stretch at
small rf (ignition limit)
University of Southern California - Department of Aerospace & Mechanical Engineering
Premixed-gas flames - stretched flames
• Counterflow configuration (Tohoku group)
• S = dU/dy – flame located where U = SL
• Increased stretch pushes flame closer to stagnation plane decreased volume of radiant products
• Similar Le effects as curved flames
• Results: Dual limits, flammability extension even for Le >1,
multiple solutions (which ones are stable?)
• Dual limits & Le effects
seen in µg experiments,
but evidence for
multivalued behavior
inconclusive
University of Southern California - Department of Aerospace & Mechanical Engineering
“FLAME BALLS”
• Zeldovich,
1944:
stationary
spherical flames possible: 2T &
2C = 0 have solutions for
unbounded domain in spherical
geometry - T ~ C1 + C2/r - bounded
as r  ∞
• Not possible for
• Cylinder (T ~ C1 + C2ln(r))
• Plane (T ~ C1+C2r)
• Mass conservation requires U  0
everywhere (no convection) – only
diffusive transport
University of Southern California - Department of Aerospace & Mechanical Engineering
“FLAME BALLS”
• T ~ 1/r - unlike propagating flame where T ~ e-r dominated by 1/r tail (with r3 volume effects!)
• Flame ball: a tiny dog wagged by an enormous tail
T*
C ~ 1-1/r
Temperature
Fuel concentration
T ~ 1/r
T•
Interior filled
with combustion
products
Fuel & oxygen
diffuse inward
Reaction zone
Heat &
products
diffuse outward
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame balls - history
• Zeldovich, 1944; Joulin,
1985; Buckmaster, 1985:
adiabatic flame balls are
unstable
• Ronney (1990): seemingly
stable, stationary flame balls
accidentally discovered H2air mixtures in drop-tower
experiment
but
test
duration very short
• Confirmed
in
parabolic
aircraft flights (Ronney et
al.,
1994)
but
g-jitter
problematic
QuickTime™ and a
Video decompressor
are needed to see this picture.
Drop tower test
QuickTime™ and a
Video decompressor
are needed to see this picture.
Aircraft test
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame balls - history
• Buckmaster, Joulin & collaborators: window of stable
conditions with radiative loss & low Lewis number
15
Dimensionless flame ball radius (R)
Uns table to 3 -d disturbances
10
Stable
Equation of curve :
-2
R ln(R) = Q
5
Uns table to 1 -d
disturbances
0
0
0.05
0.1
0.15
Dimensionless heat loss (Q)
0.2
Buckmaster, Joulin, Ronney (1990)
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame balls in space
• STS-83 & STS-94, 1997
• Stable for > 500 seconds (!)
• Very long evolution time
scales ~ (r*)2/a ≈ 100 s
• Weakest flames ever burned
(1 – 2 Watts/ball)
QuickTime™ and a
Video decompressor
are needed to see this picture.
4.9% H2- 9.8% O2 - 85.3% CO2, 500 sec
QuickTime™ and a
Video decompressor
are needed to see this picture.
4.0% H2-air, 223 sec elapsed time
QuickTime™ and a
Video decompressor
are needed to see this picture.
6.6% H2- 13.2% O2 - 79.2% SF6, 500 sec
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame ball drift
• Flame balls always drifted apart at a continually
decreasing rate
• Flame balls interact by
• (A) warming each other - attractive
• (B) depleting each other’s fuel - repulsive
• Analysis (Buckmaster & Ronney, 1998)
» Adiabatic flame balls, two effects exactly cancel
» Non-adiabatic flame balls, fuel effect wins - thermal
effect disappears at large spacings due to radiative loss
Fuel concentration
profile
Lower fuel
concentration
Higher fuel
concentration
DRIFT
DIRECTION
Affected ball
Adjacent ball
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame ball drift
10
QuickTime™ and a
Video decompressor
are needed to see this picture.
Radius of separation (cm)
4.9% H
- 9.8% O
2
2
- 85.3% CO
2
MSL-1/STS-83
3 flame balls
Space experiments
Theory (Buckmaster
& Ronney, 1998)
1
10
100
1000
T ime (seconds)
University of Southern California - Department of Aerospace & Mechanical Engineering
G-jitter effects on flame balls
• Radiometer data drastically affected by impulses caused
by small VRCS thrusters used to control Orbiter attitude
•
•
Temperature data moderately affected
Vibrations (zero integrated impulse) - no effect
• Flame balls & their surrounding hot gas fields are very
sensitive accelerometers!
• Requested & received “free drift” (no thruster firings)
during most subsequent tests with superb results
University of Southern California - Department of Aerospace & Mechanical Engineering
G-jitter effects on flame balls
QuickTime™ and a
Video decompressor
are needed to see this picture.
QuickTime™ and a
Video decompressor
are needed to see this picture.
0.2
0.15
80
Beginning of test
VCRS activities
40
0.05
20
0
0
-0.05
-0.1
-20
0
100
200
300
400
T ime from ignition (seconds)
Without free drift
500
80
0.1
Radiometer voltage (V)
0.1
Acceleration (µg)
60
60
0.05
40
0
20
-0.05
Acceleration (micro-g)
0.15
Radiometer voltage
100
Beginning of test
0
-0.1
-20
15:33:20
15:35:00
15:36:40
GMT
15:38:20
15:40:00
With free drift
University of Southern California - Department of Aerospace & Mechanical Engineering
G-jitter effects on flame balls - continued
• Flame balls seem to respond more strongly than
ballistically to acceleration impulses, I.e. change in ball
velocity ≈ 2 ∫g dt
• Consistent with “added mass” effect - maximum
possible acceleration of spherical bubble is 2g
dt (mm/s)
4
STS-94/MSL-1R, TP 13AR
7.0% H - 14.0% O - 79.0% SF
2
2
2
6
Impulse
0
2
1
Flame ball velocity
-1
0
100
200
300
400
500
(mm/s)
Impulse
g
 •
QuickTime™ and a
Video decompressor
are needed to see this picture.
z
1
3
y
Drift velocity, V
3 atm total pressure
1 flame ball
0
Time from ignition (s)
University of Southern California - Department of Aerospace & Mechanical Engineering
Comparison of predicted & measured radii
•
• Detailed
numerical modeling (Yale, USC)
•
Unsatisfactory agreement with experiment - even with
chemical models that correctly predict planar H2-air
burning velocities!
 Results sensitive to H + O2 + H2O  HO2 + H2O - not
important for planar flames away from limits
400
12
Peters & Rogg
GRI
Yetter
Experiments
Burning velocity (cm/s)
Radius (mm)
10
8
6
4
2
300
250
200
150
100
50
0
0
3
3.5
4
Mole percent H
4.5
2
in air
Yetter
GRI
Peters & Rogg
Experiments
350
5
10
100
Mole percent H
2
in air
H2-air mixtures, 1 atm
University of Southern California - Department of Aerospace & Mechanical Engineering
Chemical rate discrepancies
17
2 10
H+O +H O HO +H O
2
2
2
2
H+O O+OH
2
2
Lindstedt
/mol
2 10
11
17
1.5 10
17
GRI
Leeds
1 10
11
Williams
1 10
17
Yetter
Peters
Lindstedt
GRI
5 10
5 10
10
16
Leeds
Yetter
0
600
700
800
900
1000
1100
0 10
1200
/mol/sec)
Williams
Peters
11
3
K(H+O +H O  HO +H O) (cm
2
2
2
2
6
1.5 10
K(H+O 2  O+OH) (cm
/sec)
2.5 10
0
T emperature (Kelvins)
Competition between branching & recombination depends not
only on [M] ~ P, but also Chaperon efficiencies, esp. H2O
University of Southern California - Department of Aerospace & Mechanical Engineering
Reabsorption effects in flame balls
• Not included in radiation model but
Lplanck,CO2 ≈ 3.5 cm at 300K; Lplanck, SF6 ≈ 0.26 cm at 300K
• Decreases heat loss, widens flammability limits
• Agreement much better when CO2 & SF6 radiation ignored!
(limit of zero absorption length for CO2 & SF6)
• Still better with optically thick model (Ju et al.)
1
H 2-O 2-CO 2 mixtures (H
predictions
(optically thin, no
CO 2 radiation)
0.6
Radiative loss (Watts)
Flame ball radius (cm)
0.8
4
:O 2 = 1:2)
2
experiments
predictions
(optically thin,
with CO 2 radiation)
0.4
0.2
predictions
(optically thick)
0
4
5
3
Mole % H
7
8
2
predictions
(optically thick )
predictions
(optically thin,
w ith CO
radiation)
2
2
ex perime nts
1
0
6
predictions
(optically thin, no
CO radiation)
4
5
2
6
Mole % H
7
8
2
H2-O2-CO2 mixtures (H2:O2 = 1:2)
University of Southern California - Department of Aerospace & Mechanical Engineering
Nonpremixed-gas flames - counterflow
• Counterflow flames
• Nonpremixed flames – less
freedom of movement –
flame must lie where
stoichiometric flux ratio
maintained
• Radiating gas volume ~
flame thickness ~ (a/S)1/2
• Computations
&
µg
experiments – simple Cshaped dual-limit response
• Conductive loss to burners
at low S? (Smin)-1 ≈ tcond
CH4-N2 vs. air
(Maruta et al. 1998)
University of Southern California - Department of Aerospace & Mechanical Engineering
Nonpremixed-gas flames - gas-jet flames
• Flame height (Lf) and residence time (tjet) determined by
equating diffusion time (d2/D) to convection time (Lf/U)
• Mass conservation: U(0)d(0)2 ~ U(Lf)d(Lf)2 (round jet);
U(0)d(0) ~ U(Lf)d(Lf) (slot jet)
• Buoyant flow: U(Lf) ~ (gLf)1/2; nonbuoyant: U(Lf) = U(0)
Geometry
Flow
Lff
tjjeett
Round-jet Momentum
Uoodoo22/D
doo22/D
Round-jet
Buoyant
Uoodoo22/D
(Uoodoo22/gD)11//22
Slot-jet
Momentum
Uoodoo22/D
doo22/D
Slot-jet
Buoyant
(Uoo44doo44/D22g)11//33 (Uoo22doo22/g22D)11//33
University of Southern California - Department of Aerospace & Mechanical Engineering
Gas-jet flames - results
• Lf ≈ same at 1g or µg for round jet
100
f
L /d
o
Sund erlan d et al. (1g)
Sund erlan d et al. (µg, 1 a tm)
Sund erlan d et al. (µg, 0.5 atm )
Sund erlan d et al. (µg, 0.2 5 atm)
Coch ran a nd Mas ica (µg)
Baha dori et a l. (µg)
Baha dori & Stocke r (µg)
10
1
•
1
10
100
1000
Reynolds number (Re)
Sunderland et al. (1998) - CH4/air
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame lengths at 1g and µg
• tjet larger at µg than 1g for round jet
 Larger µg flame width ~ (Dtjet)1/2 - greater difference at low
Re due to axial diffusion & buoyancy effects
 Greater radiative loss fraction at µg (≈ 50% vs. 8%)
w/d
o
10
Sunderland et al. (1g)
Sunderland et al. (µg, 1 atm)
Sunderland et al. (µg, 0.5 atm)
Sunderland et al. (µg, 0.25 atm)
Cochran and Mas ic a
Bahadori et al.
1
10
100
1000
Reynolds number (Re)
Sunderland et al. (1998) - CH4/air
University of Southern California - Department of Aerospace & Mechanical Engineering
Turbulent flame lengths at 1g and µg
• Turbulent flames: D ~ u’LI; u’ ~ Uo; LI ~ do
 Lf ~ do (independent of Re)
• Differences between 1g & µg seen even at high Re buoyancy effects depend on entire plume
Flame length / nozzle diameter
500
Microgravity
400
300
Blow-off
limits
200
Earth gravity
100
0
0
1000
2000
3000
4000
5000
6000
7000
8000
Jet Reynolds number
Bahadori et al. (1997) - C3H8/air
University of Southern California - Department of Aerospace & Mechanical Engineering
Sooting gas-jet flames at 1g and µg
• Typically greater at µg due to larger tjet - outweighs lower T
• Smoke points seen at µg - WHY???
• tjet ~ Uo1/2 for buoyant flames BUT...
• tjet independent of Uo for nonbuoyant flames !
• Axial diffusion effects negligible at Re > 50
• Thermophoresis effects - concentrates soot in annulus
1g
µg
n-butane in air, 10mm diameter jet, Re = 42 - Fujita et al., 1997
University of Southern California - Department of Aerospace & Mechanical Engineering
EXAMPLES - Condensed-phase - droplets
• Spherically-symmetric model (Godsave, Spalding 1953)
• Steady burning possible - similar to flame balls
» (large radii: transport diffusion-dominated)
• Mass burning rate = (π/4)rdddK; K = (8/rdCP) ln(1+B)
• Flame diameter df = dd ln(1+B) / ln(1+f)
• Regressing droplet: ddo2 - dd(t)2 = Kt if quasi-steady
• 1st µg experiment - Kumagai (1957) - K(µg) < K(1g)
1.2
Droplet
(heptane in air)
0.8
QuickTime™ and a
decompressor
are needed to see this picture.
f
Normalized temperature
-T )
(T - T • ) / (T
•
1
0.6
Flame ball
Propagating flame
( d/r = 1/10)
f
0.4
0.2
0
0.1
1
10
Radius / Radius of flame
100
University of Southern California - Department of Aerospace & Mechanical Engineering
EXAMPLES - Condensed-phase - droplets
• Dual-limit behavior
• Residence-time limited (small dd): tdrop = df2/a ≤ tchem
• Heat loss (large dd): tdrop ≥ trad
• Radiative limit at large dd confirmed by µg experiments
Marchese, Dryer,
Nayagam (1999)
University of Southern California - Department of Aerospace & Mechanical Engineering
Droplet combustion - continued
• Large droplets not quasi-steady
• Extinction occurs at large dd, but dd decreases during burn quasi-steady extinction not observable
• K & df/dd not constant - depend on ddo & time
• Large time scale for diffusion of radiative products to farfield & O2 from far-field
• Soot accumulation dependent on ddo
• Absorption of H2O from products by fuel
University of Southern California - Department of Aerospace & Mechanical Engineering
Soot formation in µg droplet combustion
0 sec
0.5 sec
0.2 sec
0.6 sec
0.3 sec
0.7 sec
0.4 sec
0.8 sec
n-heptane in air (Lee et al., 1998)
University of Southern California - Department of Aerospace & Mechanical Engineering
Candle flames
• Similar to quasi-steady droplet but near-field not spherical
• Space experiments (Dietrich et al., 1994, 1997)
• Nearly hemispherical at µg
• Steady for many minutes - probably > df 2/a
• Eventual extinguishment - probably due to O2 depletion
QuickTime™ and a
Cinepak decompressor
are needed to see this picture.
1g
µg
University of Southern California - Department of Aerospace & Mechanical Engineering
Candle flames - oscillations
• Oscillations before extinguishment, except for small df
•
•
•
•
Near-limit oscillations of spherical flames? (Matalon)
Edge-flame instability? (Buckmaster)
Both models require high Le & near-extinction conditions
Some evidence in droplets also (Nayagam et al., 1998)
QuickTime™ and a
Video decompressor
are needed to see this picture.
QuickTime™ and a
decompressor
are needed to see this picture.
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame spread over solid fuel beds
• deRis (1968); Delichatsios (1986): flame spread rate (Sf)
with opposing flow U, infinite-rate kinetics (mixing limited)
Sf 
g
Tf  Tv
4 rsC p,s s Tv  T

Sf  U
rCP
srsCp,s
(thin fuel) - independent of U
T f  Tv 


Tv  T 
2
(thick fuel) - Sf ~ U1
• Diffusive transport time scale (tdiff) ≈ d/U ≈ a/U2
• Heat loss parameter H ~ tdiff/trad = a/U2trad
University of Southern California - Department of Aerospace & Mechanical Engineering
Condensed-phase combustion - flame spread
• Sf lower at µg: U = Sf << U(1g)  higher H
• Infinite-rate kinetics limit not achieved at 21% O2 !
• Dual-limit behavior
•
•
•
•
Residence-time limited (large U): tdiff ≤ tchem
Heat loss (small U): tdiff ≥ trad
Confirmed by thin-fuel µg experiments (Olson et al., 1990)
Most robust U ≈ 10 cm/s - less than 1g buoyant flow!
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame spread - continued
• Radiation not all lost if ambient atmosphere absorbs;
Honda & Ronney, 1998:
• O2-N2, O2-He, O2-Ar: Sf(1g) > Sf(µg) due to radiative loss
• O2-CO2, O2-SF6: Sf(1g) < Sf(µg) due to reabsorption
• International Space Station uses CO2 fire extinguishers!
5
4
2.5
µg
1g
1g (CO
2
µg (CO
1g (SF
S f (cm/sec)
S f (cm/sec)
µg (SF
3
2
2
6
6
)
)
)
)
1.5
2
1
0.5
1
O -He atmospheres
2
0
15
20
25
30
O concentration (mole % )
2
35
40
0
20
25
30
35
40
O concentration (mole %)
45
50
2
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame spread - continued
• All fronts thicker at µg (d ≈ a/U)
• With reabsorption, difference between 1g and µg is larger
QuickTime™ and a
Video decompressor
are needed to see this picture.
19% O2 in N2
(optically thin)
QuickTime™ and a
Video decompressor
are needed to see this picture.
42% O2 in SF6
(optically thick)
University of Southern California - Department of Aerospace & Mechanical Engineering
Flame spread - thick fuels
• Steady Sf not possible at µg
since Sf ~ U - chicken & egg
problem
2

T

T
rCP  f v 

Sf  U
srsCp,s Tv  T 
QuickTi me™ and a
V ideo decompressor
are needed to see t his pict ure.
• Instead Sf ~ t-1/2, decreases until
extinction (Altenkirch et al., 1996)
• Strongly radiating flames in O2CO2 or O2-SF6 could spread
steadily at µg (egg = radiation)
1/ 2


L ag2
S f  

a
r
C

T

T


T

T




g
s
P,s
s
v

g
f
v


• Evidence seen recently (Honda,
Son & Ronney, 2001) using foam
fuels (lower srsCP,s  faster Sf)
University of Southern California - Department of Aerospace & Mechanical Engineering
Fingering flame spread at µg
• Olson et al. 1998 - space experiments
• Strong forced flow - smooth fronts, similar to 1g
• Weak or no forced flow - fingering fronts
• Radiative or conductive loss: gas-phase heat transfer lost;
heat transport through solid phase; O2 transport can only
occur through gas phase
• Two Lewis numbers?
» High U: heat transport
in gas phase; Leeff =
agas/DO2 ≈ 1
» U  0: heat transport
through solid; Leeff =
asolid/DO2 << 1

Air
6.5 cm/s
University of Southern California - Department of Aerospace & Mechanical Engineering
Summary - what have we learned?
• Time scales
• when buoyancy, radiation, etc. is important
• Radiative loss – gas-phase & soot
• causes many of the observed effects on burning rates &
extinction conditions
• double-edged sword - optically thin vs. reabsorbing
• Dual limits (high-speed blow-off & low-speed radiative)
• seen for practically all types of flames studied to date
• Spherical flames (flame balls, droplets, ≈ candle flames)
• long time scales, large domains of influence, radiative loss
• Oscillations near extinction
• common, not yet fully understood
• Chemistry
• different reactions rate-limiting for very weak flames
University of Southern California - Department of Aerospace & Mechanical Engineering
Challenges for future work
• Radiative reabsorption effects
• Apparently seen in many µg flames
• Relevant to IC engines, large furnaces, EGR, flue-gas
recirculation (d ≈ aP-1)
• Need faster computational models of radiative transport!
• High-pressure combustion
•
•
•
•
Buoyancy effects (tchem/tvis) increase with P for weak mixtures
Reabsorption effects increase with P
Turbulence more problematic
Few µg studies - mostly droplets
• 3-d effects
• Flame spread - effects of fuel bed width
• Flame balls - breakup of balls
• Spherical diffusion flames - porous sphere experiment
University of Southern California - Department of Aerospace & Mechanical Engineering
Thanks to…..
Angel Abbud-Madrid, Tom Avedisian, Yousef Bahadori, John
Buckmaster, Mun Choi, Dan Dietrich, Ed Dreizin, Fred Dryer,
Said Elgobashi, Gerard Faeth, Osamu Fujita, Guy Joulin,
Yiguang Ju, Kaoru Maruta, Moshe Matalon, John Moore,
Vedha Nayagam, Takashi Niioka, Sandra Olson, Howard
Ross, Kurt Sacksteder, Dennis Stocker, Peter Sunderland,
Gregory Sivashinsky, James Tien, Arvind Varma, Karen
Weiland, Forman Williams, Ming-Shin Wu
Institute of Mechanics, Chinese Academy of Sciences
….. and especially NASA-Glenn!!!
University of Southern California - Department of Aerospace & Mechanical Engineering