Radiation-driven Flame Spread Over Thermally-thick Fuels In Quiescent Microgravity Environments

Radiation-Driven Flame Spread Over Thermally-Thick Fuels in Quiescent Microgravity Environments Youngjin Son & Paul D. Ronney Department of Aerospace & Mechanical Engineering University of Southern California Los Angeles, CA 90089-1453 USA Twenty Ninth International Symposium on Combustion Sapporo, Japan, July 22 – July 26, 2002 Supported by the NASA Glenn Research Center under grant NCC3-671

Outline 

Prior results  

  

Objectives Apparatus Results  



Thin fuels, esp. effects of radiation at µg Thick fuels

Spread rates Radiative fluxes

Conclusions

University of Southern California - Department of Aerospace & Mechanical

Theoretical Background – Thin Fuels Ideal flame spread rate (Sf) without radiation is steady Ideal for thin fuels even at µg -> Sf independent of opposed flow velocity (U) (deRis, 1969)

λg Sf = ρ s C p ,sτ s

 T f − Tv     Tv − T∞ 

University of Southern California - Department of Aerospace & Mechanical

Current understanding (continued)  Experiments - thermally-thin fuels (Honda & Ronney, 1998) Sf at  Radiatively non-participating He, Ar, N2 diluents: 1g > Sf at µg, minimum O2 concentration lower at 1g vs. µg  Opposite for radiatively participating CO2, SF6 diluents

 Behavior for non-radiating diluents attributed to radiative loss - µg flames thicker, more volume  Behavior for radiating diluents attributed to  Reabsorption of emitted radiation (reduced heat loss)  Re-radiation to surface (increased Sf)

University of Southern California - Department of Aerospace & Mechanical

Thin Fuel Results (continued)

f

S (cm/s)

3

1 0.8 0.6

1g-SF

0.4

ug-CO

0.2 15

6

1g SF6 ug-SF µg SF66 1g -He 1g He µg He ug -He 1g CO2 1g-CO 2 µg CO2 20

25

30

35

40

45

50

2

55

O concentration (mole %) 2

University of Southern California - Department of Aerospace & Mechanical

Current understanding (continued)  Experiments - thermally-thin fuels (Honda & Ronney, 1998) Sf at  Radiatively non-participating He, Ar, N2 diluents: 1g > Sf at µg, minimum O2 concentration lower at 1g vs. µg  Opposite for radiatively participating CO2, SF6 diluents

 Behavior for non-radiating diluents attributed to radiative loss - µg flames thicker, more volume, more loss  Behavior for radiating diluents attributed to  Reabsorption of emitted radiation (reduced heat loss)  Re-radiation to surface (increased Sf)

University of Southern California - Department of Aerospace & Mechanical

Schematic of radiation & reabsorption Absorption & Re-radiation (CO2 or SF6) Radiation from flame Oxygen

U Flame

Fuel

Fuel Bed University of Southern California - Department of Aerospace & Mechanical

Theory of thick fuel flame spread  In contrast to thin fuels, Sf dependent on opposed flow velocity (deRis, 1969) 2   λ g ρ g Cp, g Tf − Tv   Sf = U λ s ρ sCp,s  Tv − T∞ 

 Conventional wisdom: steady Sf not possible at µg without forced flow since Sf ~ U  Instead Sf ~ t-1/2 , decreases until extinction (Altenkirch et al., 1996, 1998)  At 1g, buoyant flow provides U - steady spread possible University of Southern California - Department of Aerospace & Mechanical

Hypothesis  de Ris (1969): Radiative transfer from external source to fuel bed leads to steady spread over thick fuel bed even if U = 0

qδ 2

Sf =

ρsCP ,sλs(Tv − T∞ )

2

q = radiative flux per unit area, δ = length of radiating zone

 … but the hot gases also radiate, especially in O2-CO2 & O2-SF6 atmospheres  Estimation of radiative flux from flame to fuel bed q ~ Λ δ ~ Λ (α /Sf) leads to:

 2  1/ 2 Λα g   Sf =   α ρ C λ T − T − λ T − T ( ) g( f v )  g s P ,s s v ∞ Λ

= radiative emission per unit volume (=∇ • qr )

University of Southern California - Department of Aerospace & Mechanical

Objectives  Test the hypothesis that flame spread over thick fuel beds can be steady at quiescent µg conditions due to radiative transfer from flame to fuel surface  Measure spread rates as a function of  Gravity level (1g or µg)  O2 concentration  Pressure  Fuel sample thickness

 Measure radiative emissions from gas & solid

University of Southern California - Department of Aerospace & Mechanical

Approach  µg experiments in 2.2 second drop tower  Problem with conventional thick fuels  Low Sf (e.g. PMMA): » Time scale ~ α / Sf2 too large for drop towers » Length scale ~ α / Sf possibly too large even in space  Need very low ρ sλ sCp,s - use foams  Also high pressure - ρ g higher, Λ higher

 Fuel  Polyphenolic Foam (used in floral arranging)  Density : 0.0290 g/cm3  1 sided and 2 sided spread

 Measurements  Imaging via direct video & shearing interferometry  Radiometers University of Southern California - Department of Aerospace & Mechanical

Fuel sample configuration  Ignition via nitrocellulose membrane  Interferometer to image changes in gas density (side view)  Direct video (front view)  Spot radiometers

Kanthal hotwire Hole

Nitrocellulose Membrane Camera

Radiometers Interferometer Field of view Fuel

Side View

Front View

University of Southern California - Department of Aerospace & Mechanical

Images at 1g and µg Front View

Side View

µg test

1g test

40% O2 in CO2 @ 4 atm, polyphenolic foam, density = 0.027 g/cm3 Thicker flame at µg (δ ~ α / U, U small at µg - no buoyant flow) University of Southern California - Department of Aerospace & Mechanical

Flame spread rate determination Relative flam e position (cm )

8

 

ug 1g

7 6 5

Polyphenolic foam

4

0.0267 g/cm 40%O -CO

3 2 1.2

2

3

2

4atm total pressure 1.4

1.6

1.8 Tim e(sec)

2

2.2

2.4

Steady Sf possible at µg With foam fuel, can reach steady Sf even in 2.2 sec drop tower test

University of Southern California - Department of Aerospace & Mechanical

Flame spread vs. O2 concentration Spread rate (cm/sec)

10

ug CO ug N

1

2

1g CO 1g N

20

25

30

35

Mole percent O

40

45

2

2

2

50

2

 For CO2, Sf at µg is higher than at 1g, especially at low O2 concentrations, whereas for N2, µg and 1g are similar  At µg, Sf can be higher in CO2 than N2 at the same % O2  For CO2 but not N2, the minimum O2 concentration supporting combustion is lower at µg University of Southern California - Department of Aerospace & Mechanical

University of Southern California - Department of Aerospace & Mechanical

Spread rate (cm/sec)

Flame Spread vs. Pressure 7 5

40% O - 60% diluent

3

0.0267 g/cm

2

polyphenolic foam 3

ug CO

1 0.8 0.6

ug N

0.2 0.2

2

1g CO

0.4

1g N

0.4

0.6 0.8 1

3

5

2

2

2

7

Pressure (atm )

 For N2, Sf (µg) << Sf (1g) at low P, but for CO2, Sf (µg) ≈ Sf (1g)  Reabsorption effects more important at high P - shorter absorption lengths - allows Sf (µg) > Sf (1g)  Low P: less reabsorption, loss, Sof < Sf (1g) University of Southern Californiamore - Department Aerospace & Mechanical f (µg)

Flame Spread vs. Pressure ug CO 2 Est. (1500K) Est. (1800K)

Spread rate(cm/sec)

10

 2  1/ 2 Λα g   Sf =   α ρ C λ T − T − λ T − T ( ) ( ) g s P ,s s v ∞ g f v  

1

0

1

2

3

4

5

6

7

Pressure (atm )

 Model with no adjustable parameters reasonably consistent with experiments except at Low pressures - radiative heat loss High pressures - optically thick (factors not considered in simple model) University of Southern California - Department of Aerospace & Mechanical

Flame spread rate vs. thickness ug, 1-side ug, 2-sides

Spread rate (cm/sec)

6

1g, 1-side 1g, 2-sides

5 4 3 2 3

1 0

Polyphenolic foam , 0.0267g/cm 40%O -CO , 4atm total pressure 2

0.1

2

Thickness (cm )

1

 Sf is independent of thickness (τ ) when τ > 2 mm (thermally-thick behavior)  Thermally-thin behavior at τ < 2 mm (Sf is dependent on τ )  For thinnest samples, Sf (1-sided) ≈ 1/2 of Sf (2-sided) - consistent with the simple thermal model for thin fuels  …but trend NOT monotonic! University of Southern California - Department of Aerospace & Mechanical

CO2 vs. He diluent Spread Rate (cm /sec)

10

polyphenolic foam 4 atm total pressure 0.027g/cm

3

ug-CO2 1

ug-He 1g-C O2 1g-H e 20

25

30

35

40

Mole percent O

45

50

55

2

 CO2 much better than helium at 1g  He may be better inerting agent at µg Better efficacy per mole (⇒ storage bottle mass & volume) Much better per unit mass No physiological impact University of Southern California - Department of Aerospace & Mechanical

Radiometer configuration Flame

Radiation from flame

Re-radiation (CO2 & SF6)

Front-side radiometers (2) (A) Views hole - outward gas radiation only (B) Offset horizontally from hole - outward gas + solid radiation

Hole

Rear-side radiometer (1) (C) Views through hole measures incident gas radiation only

Fuel bed

University of Southern California - Department of Aerospace & Mechanical

Radiation (CO2 diluent, µg) 2

Radiative Flux (W /cm )

0.006 Front, Gas only Front, Gas + Surface Rear, Gas only

0.005 0.004

Blue: gas-phase radiative loss only

End of Drop

0.003

Red: gas+surface radiative loss

0.002

Green: gas-phase radiation to surface

0.001 0 -0.001 0

0.5

1 1.5 Elapsed Tim e (sec)

2

2.5

 Radiation from front & rear gas-phase radiometers show similar intensity and timing - substantial re-absorption and re-radiation  Surface radiation > gas-phase; peak is later (after flame passage)  Substantial radiative flux to fuel bed - accelerates spread University of Southern California - Department of Aerospace & Mechanical

Radiation (N2 diluent, µg) Front, G as only Front, G as + Surface Rear, Gas only

0.005

2

Radiative Flux (W /cm )

0.006

0.004

Blue: gas-phase radiative loss only

E nd of Drop

0.003

Red: gas+surface radiative loss

0.002

Green: gas-phase radiation to surface

0.001 0 -0.001 0

0.5

1

1.5

2

2.5

Elapsed Tim e (sec)

 Radiation to rear-side radiometer small compared with CO2 diluent - little importance of gas-phase radiation to fuel surface  Gas-phase loss significant - higher than CO2 - less reabsorption  Peak surface radiative loss similar to CO2 University of Southern California - Department of Aerospace & Mechanical

Radiation (CO2 diluent, 1g) Front, G as only Front, G as + Surface Rear, G as only

2

Radiative Flux (W /cm )

0.007 0.006 0.005

Blue: gas-phase radiative loss only

0.004

Red: gas+surface radiative loss

0.003 0.002

Green: gas-phase radiation to surface

0.001 0 -0.001

0

1

2 3 4 Elapsed Tim e (sec)

5

Gas-phase radiative loss less than µg case due to thinner front (less volume) Negligible re-radiation to surface Surface radiative loss similar to µg

University of Southern California - Department of Aerospace & Mechanical

Radiation (N2 diluent, 1g) Front, G as on ly Front, Gas+Surface Rear, G as only

2

Radiative Flux (W /cm )

0.007 0.006 0.005

Blue: gas-phase radiative loss only

0.004

Red: gas+surface radiative loss

0.003 0.002

Green: gas-phase radiation to surface

0.001 0 -0.001

0

0.5 1 1.5 Elapsed Tim e (sec)

2

Gas-phase radiative loss less than µg case due to thinner front (less volume) Negligible re-radiation to surface Surface radiative loss similar to µg

University of Southern California - Department of Aerospace & Mechanical

Conclusion  Drop-tower experiments show  Steady spread observed over thick fuels in quiescent µg conditions  Sometimes faster than 1g spread!

 Results due to  More significant reabsorption of emitted radiation in O2-CO2 vs. O2-N2  Thicker flames (more volume, more radiative flux & more reabsorption) at µg  Radiometer data consistent with these hypotheses

 Findings may be applicable for spacecraft fire safety  ISS uses CO2 fire extinguishers, but flames may spread faster at µg with CO2 diluent than N2 diluent due to radiative preheating of fuel!  He is a better inerting agent for the conditions tested University of Southern California - Department of Aerospace & Mechanical

Conclusion (extra)  Only O2-CO2 at µ g case where the rear radiometer shows comparable intensity and timing with two front radiometers – this means only in this case, there is substantial emission, absorption and re-emission, which is the only means to obtain radiative flux to the back-side radiometer  Radiative preheating of the fuel bed by the gas is significant in radiatively-active atmospheres at µg  Reabsorption effects can prevent massive heat losses (thus extinction) in radiatively-active atmospheres at µg  These effects are less important at 1g due to substantial U caused by buoyancy which leads to smaller flame thicknesses thus less volume of radiating gas

University of Southern California - Department of Aerospace & Mechanical

Acknowledgments (Extra) This work was supported by the NASA Glenn Research Center under grants NCC3-671. The author is grateful to Dr. Suleyman Gokoglu, Dr. Linton Honda for many helpful discussions and technical support. I also thank the FEANICS team and the NASA-Glenn 2.2-second drop tower staff for their help in coordinating and supporting the µg experiments.

University of Southern California - Department of Aerospace & Mechanical

Experimental Apparatus (Extra)  Chamber  A 20 liter chamber is used  One lexan window for the camera and two quartz window for the laser path

 lmaging system  1 CCD camera for front view  1 CCD camera for laser interferograms from side view

 Measurements  3 radiometers (details follow)  4 thermocouples (details follow)

University of Southern California - Department of Aerospace & Mechanical

Spread Rate (cm /sec)

Flame Spread vs. Density (Extra) ug 1g Fit (1g) 1 0.8

Power law fit: slope = -1.32

0.6 0.4

0.2 0.02

0.04

0.06 0.08 0.1

Density (g/cm

3

)

 As expected, Sf (µg) decreases with increasing fuel bed density for both µg and 1g University of Southern California - Department of Aerospace & Mechanical

Flame spread rate vs. thickness (extra)  Sf is independent of thickness for sufficiently thick fuel beds  Sf decreases with decreasing thickness, which is contrary to thin-fuel theory without radiation, however with radiation, the thin fuel Sf (equation below) can be lower than the thick-fuel Sf, leading to non-monotonic effects of τ s on Sf  This behavior can occur with or without an imposed or buoyant flow U, so can happen both µ g and 1g

3 2 3    S f   2U  Sf        Sf S U U f , rad   + − 1  +  − 2   − U 2 +    = 0 S f , o  S f ,o  S f ,o   S f ,o  Sf ,o    S f ,o    S f ,o  

University of Southern California - Department of Aerospace & Mechanical

Experimental Apparatus – Drop Rig Drop Frame

Window

Window

Mirror

Shearing Plate

Digital Image Processing System

Camera Mirror

Beam Laser Expander Mirror

Diffuser

≈

Test Chamber

Fiber-optic Link

VCR

Side View University of Southern California - Department of Aerospace & Mechanical

Movies - µg flame spread (extra) 40% O2-CO2 @ 4atm

40% O2-N2 @ 4atm

University of Southern California - Department of Aerospace & Mechanical