influence of flame characteristics on gas- phase fire

Influence of Flame Characteristics on Gas-Phase Fire Retardant Effectiveness

Gregory Linteris, Nicolas Bouvet*, Valeri BabushokFire Science Division, Engineering Laboratory, NIST

Fumiaki TakahashiCase Western Reserve University / NASA Glenn

Viswanath KattaWright-Patterson Air Force Base / ISSI

Stanislav Stoliarov, Peter SunderlandDept. of Fire Protection Engineering, University of Maryland

* Financial support for NB from BASF

Acknowledgements:

BASF (support of Dr. Bouvet)

Past Support:NASANISTBoeingHonweywellDupontArmy, Navy, Air ForceNRCvon Humboldt Foundation)

Other People: Harsha Chelliah, UVa; Valeri Babushok, Wing Tsang, Kevin McGrattan, Don Burgess, Mike Zachariah, Marc Rumminger, Dirk Reinelt, Greg Fallon, Ivan, Tony, Cindy, Arnold, Newton, Tania, Ian, Lloyd, NIST;

Outline:

Part I: Super-effective agents in co-flow diffusion flames

Background- flame types- iron inhibition- particles

Phosphorus- status

Part II: Flame Properties Affecting Inhibition EffectivenessFlame Base CharacteristicsRandom ThoughtsClosing Thoughts

0.4

0.6

0.8

1

0 2000 4000 6000 8000 10000

Inhibitor Volume Fraction (L/L)

CF3Br

Fe(CO)5

= 1.0

Tin = 353 K(except Sn)

CO2

X = 0.21O2,ox

Nor

mal

ized

Bur

ning

Vel

ocity

Gas-phase chemical inhibitor effectiveness varies dramatically.

* From Linteris, G.T., Rumminger, M.D., Babushok, V.I. “Catalytic Inhibition of Laminar Flames by Metal Compounds,” Progress in Energy and Combustion Science, 34:288 – 329, 2008.

0.4

0.6

0.8

1

0 2000 4000 6000 8000 10000

Inhibitor (ppm)

Norm

aliz

ed B

urni

ng V

eloc

ity

CF3BrSn(CH3)4

Fe(CO)5

MMT

= 1.0

SnCl4

Tin = 353 K(except Sn)

CO2

X = 0.21O2,ox

SbCl3



Premixed flameDiffusion flame

(Couterflow)

Diffusion flame (Coflow)

Fe(CO)5 : 540 100 0DMMP : 141 17 2CF3Br : 8 5 7

Effectiveness relative to CO2 :

Why?


Pool Fire (from: http://www.me.uwaterloo.ca/~eweckman/fire/firehome.htm

How do we study flame inhibitors?

Too difficult to extract fundamental information:=> can’t include detailed chemistry=> grid resolution is coarse=> turbulence too complex

Premixed Flames Diffusion Flames

flame

oxidizer

Counter-flow

fuelChimney

Air + Inhibitor

Fuel

Co-flow

1-DSteadyEasy to model, well developed theoryTypically used by researchersProvide fundamental reaction rate data

2-DUnsteadyHarder to model, active area of researchWidely used in codes and stds.Has many features like real fires

Two basic types of flames :

Fuel + air+ inhibitor

Nozzle Burner

1. Experiments: - all three fundamental flames, - over wide range of controlling parameters.

2. Detailed numerical simulation- with detailed kinetic mechanisms.

3. At NIST, we - build gas-phase kinetic models,- use other people’s flame codes

Goals: 1. Understand interaction of inhibitors with flame structure.

2. Eventually model:- standard tests, - full-scale fires.

Approach

Iron is super-effective relative to CF3Br in some flames,

0.4

0.6

0.8

1

0 2000 4000 6000 8000

Inhibitor %

Flam

e St

reng

th

CF3Br

Fe(CO)5

CO2

0.80.40.20.0 0.6

Premixed flame

0.05

0.10

0.15

0 0.2 0.4 0.6 0.8

Inhibitor (%)

Flam

e St

reng

th

CF3Br

Fe(CO)5

CO2

Cup-burner flame

Use detailed modeling of the cup-burner flame to find out why.

but in others, it’s relatively ineffective. Why?


Direct Numerical Simulations for Cup-Burner Flames

• Time-dependent 2D governing equations mass, momentum, species, energy conservation, & state equations,

including a body-force term optically thin-media radiative heat losses from CO2, H2O, CH4,and CO

• Variable thermochemical/transport properties hi from polynomial curve-fits , , and D from molecular dynamics and mixture rules

• Detailed reaction mechanism (GRI-Mech v1.2) 31 species/346 reactions for CH4-O2 combustion + inert (He, Ar) 82 species/1520 reactions for CH4-O2 – F system

Calculated temperature contours for one oscillation cycle

Adding slightly more agent causes flame to liftoff

a.) b.)

CH4-air flame, Xi < Xi|critical Xi > Xi|critical

Can predict critical volume fraction of agent for blow-off well.

OS

F

OS

Fuel 0

0.1

0.2

0.3

0.4

Br2 CF3Br CF3H Ar He N2 CO2 CO2(mg)

Volu

me

Frac

tion

for E

xtin

ctio

n

experiment

prediction

Air + inhibitor

Agent

Review: Radical reactions dominate combustion

- radical attack dominates fuel decomposition

- radical chain branching creates the radicals

- CO consumption:

H H H

H – C – C – C – H

H H H

H H H

H – C – C – C – H

H H• H + H – H

•

H + O2 OH + O• • •

CO + OH CO2 + H• •

- because of chain branching, often have super-equilibrium of radicals

How does iron inhibit flames?

FeO + H2O = Fe(OH)2Fe(OH)2 + H = FeOHFeOH + H = FeO------------------------------------------------Net: H + H = H2

Fe(OH)2

FeO

FeOH

+H

+H

+H2O

Fe(CO)5

Fe

FeO2

+O

+O2+M

Catalytic cycle

=> Gas-phase iron species recombine radicals.

0.00

0.05

0.10

0.15

0 500Fe(CO)5 Volume Fraction in Air (ppm)

Flam

e St

reng

th

0

1

2

3

4

Max

ium

u Q

vv x

105 1

/(cm

-sr)

Calc

Expt.

Predicted and measured inhibition don’t agree. => This was a surprise. It was expected to work.

From: Linteris, G.T., Katta, V.R., and Takahashi, F., “Experimental and Numerical Evaluation of Metallic Compounds for Suppressing Cup-Burner Flames,” Combustion and Flame, 138:78-96, 2004.

Iron species vapor pressures are strong f(T).

1.E-10

1.E-09

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

500 1000 1500 2000

Temperature K

Gas

-Pha

se V

olum

e Fr

actio

n

FeO

Fe

Fe(OH)2

=> potential for condensation.


Use model to determine flame structure. Check for super-saturation of Fe compounds:

=> Particles: sink for active Fespecies

=> [FeO]/[FeO]sat = 103 to 105 => condensation is likely

=> Temperature at reaction kernel is low.

1.00E-10

1.00E-09

1.00E-08

1.00E-07

1.00E-06

1.00E-05

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

2 4 6 8 10 12Radial Distance (mm)

Mol

e Fr

actio

n

400

600

800

1000

1200

1400

1600

Tem

pera

ture

(K)

XFeO

XFeO,sat

CH4O2T

10-10

10-8

100

10-6

10-4

10-2

Fe(OH)2

FeO

FeOH

+H

+H

+H2O

Catalytic cycle

Check for particles with laser scattering system

Lock-in Amp

Computer

Pre-amp L

Laser

C

M

PMT

Pre-amp

Po

G

Burner

PMT

M

F P L

S

Po

PMT

F

Lock-in Amp

IS

Lock-in Amp

Fiber

Pre-amp

LL

Bellows

Chimney

F

-20

-18

-15

-13

-11 -8 -6 -3 -1 1 4 6 9 11 13 16 18

0

3

6

9

12

15

18

-20 -15 -10 -5 0 5 10 15Radial Position (cm)

0

5

10

15

20

Ht. A

bove

Bur

ner

(cm

)

0 L/L

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49

S1

S4

S7

S10

S13

S16

S19

S22

-20 -15 -10 -5 0 5 10 15Radial Position (cm )

0

5

10

15

20

Ht.

Abo

ve B

urne

r (c

m)

100 L/L

-20

-18

-15

-13

-11 -8 -6 -3 -1 1 4 6 9 11 13 16 18

0

3

6

9

12

15

18

-20 -15 -10 -5 0 5 10 15

Radial Position (cm)

0

5

10

15

20

Ht.

Abo

ve B

urne

r (c

m)

200 L/L

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49

S1

S4

S7

S10

S13

S16

S19

S22


0

5

10

15

20

Ht.

Abo

ve B

urne

r (c

m)

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49

S1

S4

S7

S10

S13

S16

S19

S22


0

5

10

15

20

Ht.

Abo

ve B

urne

r (c

m)

450 L/L

325 L/L

Particle scattering increases with XFe(CO)5 in air stream.

0.00

0.05

0.10

0.15

0 500Fe(CO)5 Volume Fraction in Air (ppm)

Flam

e St

rneg

th

0

1

2

3

4

Part

icle

Sca

tterin

g C

ross

-sec

tion

Calc

Expt.Fl

ame

Stre

ngth


Scattering from particles is correlated with low effectiveness

Iron as a fire suppressant?

1. Much less effective than expected.

2. Gas-phase catalytic reactions => lower the radical concentrations. But…

3. Particles are a sink for the active species.

4. Lower temperature in the stabilization region leads to more particle formation.

What about DMMP?

Conclusions (for iron in cup-burner flames):

Cup Burner Extinction with DMMP and CO2 , (methane-air, 70°C)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0 2000 4000 6000 8000 10000

CO2Vo

lume Fractio

n in Oxidizer

DMMP Volume Fraction In Oxidizer (uL/L)

Exctinction of CH4‐Air Cup Burner flame, T=70 oC

No Flame

Flame

From: G.T. Linteris, N. Bouvet, V.I. Babushok, F. Takahashi, V.R. Katta,Experimental and numerical simulations of the gas-phase effectiveness of phosphorus compounds, in Fire and Materials 2015, 14th international conference, Interscience Communications, London, UK, 2 - 4 February, 2015.

0

2

4

6

8

10

12

14

16

18

0 0.5 1 1.5

CO2Vo

lume Fractio

n in Oxidizer

DMMP Volume Fraction In Oxidizer (uL/L)

Exctinction of CH4‐Air Cup Burner Flame, T=70 oC

Br2

DMMP

Phosphorus vs. Bromine in Cup Burner

DMMP about 4 x better than Br at low conc.

But loses effectiveness at higher conc.

%

Scattering measurements here

With DMMP – 11 mm above burner

Rel

ativ

eR

ayle

igh

Scat

.Sig

nal

Rayleigh scattering in flame with added DMMP in air ( 1000 ppm)

Uncertainties discussed in: G.T. Linteris, N. Bouvet, V.I. Babushok, F. Takahashi, V.R. Katta,Experimental and numerical simulations of the gas-phase effectiveness of phosphorus compounds, in Fire and Materials 2015, 14th international conference, InterscienceCommunications, London, UK, 2 - 4 February, 2015.

Rel

ativ

eR

ayle

igh

Scat

.Sig

nal

Ht. above burner

Rayleigh scattering in flame with added DMMP in air ( 1000 ppm)

Uncertainties discussed in: G.T. Linteris, N. Bouvet, V.I. Babushok, F. Takahashi, V.R. Katta,Experimental and numerical simulations of the gas-phase effectiveness of phosphorus compounds, in Fire and Materials 2015, 14th international conference, InterscienceCommunications, London, UK, 2 - 4 February, 2015.

1. DMMP added to cup-burner flames does lose its effectiveness dramatically.

2. Particles do form – but loss of effectivenesscould also be due to the HC component.

3. Further experiments and modeling should determine why.

Current Status

=> Influence of cup-burner flame properties on inhibitor effectiveness.

Part II

Flame is stabilized at the base, where:

- Temperature is lower.

- Mixing is good, so reactants tend to be premixed (due to flame lifting at edge, and entrainment).

- Flame oscillates, making the flame easier to extinguish.

Flame Flicker: makes flame easier to extinguish

2.8 cm

CH484.1% Air + 15.9% CO2

UCH4= 0.92 cm/s, Uox= 10.7 cm/s, XCO2 = 0.14t = 0 s

Flame flicker: 1. Makes flame easier to

extinguish.2. Without flicker, need 30%

more agent for extinguishment.

3. Some agents enhance flicker, some retard it.

4. Changes with agent loading.

Flame Structure (Showing Flicker) t=0.00 s

Flame Structure (Showing Flicker) t=0.08 s

UCH4= 0.92 cm/s, Uox= 10.7 cm/s, XCO2 = 0.14t = 0.08 s

Flame flicker: 1. Makes flame easier to

extinguish.2. Without flicker, need 30%

more agent for extinguishment.

3. Some agents enhance flicker, some retard it.

4. Changes with agent loading.

Flame lift-off

CH4 - Air + 1.66 % Br2Ch4 – Air + 2.46% CF3Br

Base is:- lower temperature (1400 K) as compared to higher up in the flame (1850 K).- Lifted, more premixed, with

+ more radical super equilibrium,+ better inhibition at base, so blows off there first + different chemsitry, due to different reactants: e.g., better regeneration of HBr

-6E-05

-5E-05

-4E-05

-3E-05

-2E-05

-1E-05

0

1E-05

6 7 8 9 10 110

500

1000

1500

2000

CH4+Br= CH3+HBr

CH2O+Br=HCO+HBr

Br+Br+M=Br2+M

1

-4

0

F+

F-

Fnet

T

Trailing flame, 2.46 % CF3Br

-5

HBr+OH=Br+H2O

HBr+H=Br + H2

Br2+H=Br+HBr

-3

-2

-1

-6Sum

of R

eact

ion

Rat

es m

ole

cm s

(x

10

)-3

-14

-6.00E-05

-5.00E-05

-4.00E-05

-3.00E-05

-2.00E-05

-1.00E-05

0.00E+00

1.00E-05

6 7 8 9 10 110

500

1000

1500

2000

CH4+Br= CH3+HBr

CH2O+Br=HCO+HBr

Br+Br+M=Br2+M

F+

F-

Fnet

T

Trailing flame, 1.66 % Br2

HBr+OH=Br+H2O

HBr+H=Br + H2

Br2+H=Br+HBr

T (K

) or B

r Rea

ctio

n Fl

ux m

ole

cm s

(x

10

)-3

-14

-5E-04

-4E-04

-3E-04

-2E-04

-1E-04

0

1E-04

2E-04

6 7 8 9 10 110

500

1000

1500

2000

F+

F-

Fnet

T

Reaction Kernel, 2.46 % CF3Br

HBr+OH=Br+H2O

HBr+H=Br + H2

Br2+H => Br + HBr

10

-40

0

-50

-30

-20

-10

20

C2H6+Br=C2H5+HBr

CH4+Br=CH3+HBr

CH2O+Br=HCO+HBr

r (mm)

Sum

of R

eact

ion

Rat

es m

ole

cm s

(x

10

)-3

-14

-5E-04

-4E-04

-3E-04

-2E-04

-1E-04

0

1E-04

2E-04

6 7 8 9 10 110

500

1000

1500

2000

F+

F-

Fnet

T

Reaction Kernel, 1.66 % Br2

HBr+OH=Br+H2O

HBr+H=Br + H2

Br2+H => Br + HBr

C2H6+Br=C2H5+HBr

CH4+Br=CH3+HBr

CH2O+Br=HCO+HBr

r (mm)

T (K

) or B

r Rea

ctio

n Fl

ux m

ole

cm s

(x

10

)-3

-14

Influence of Flame Base on Inhibition Chemistry

10

500

1000

1500

2000

T (K

) or B

r Rea

ctio

n Fl

ux m

ole

cm s

(x

10

)-3

-14

10

500

1000

1500

2000

T (K

) or B

r Rea

ctio

n Fl

ux m

ole

cm s

(x

10

)-3

-14

Burner

Inhibiting reactions 10x higher in base

Because HBr regeneration steps are higher in base

Other Flame Properties Influencing Inhibitor Effectiveness

(Random thoughts)

0.0

0.2

0.4

0.6

0.8

1.0

0 100 200 300 400 500Fe(CO)5 Mole Fraction (ppm)

Nor

mal

ized

Bur

ning

Vel

ocity

= 1.0

Different flame metrics influence effectiveness:=> Xa for 30% reduction in flame speed would imply high effectiveness,

X = 0.212 ,oxO

0.0

0.2

0.4

0.6

0.8

1.0

0 100 200 300 400 500Fe(CO)5 Mole Fraction (ppm)

Nor

mal

ized

Bur

ning

Vel

ocity

X = 0.21O2 ,ox

= 1.0

=> but, Xa for 5 cm/s flame speed would imply low effectiveness,

5 cm/s

Different behavior sometimes shows up at higher concentrations.

1

10

100

1000

CO2 CF3H CF3Br TMT MMT Fe(CO)5

Perf

orm

ance

Rel

ativ

e to

CO

2

Agent

a=430

a=490

a=550

30

45

Cup, pure

Cup, 10% less CO2(like LOI)

Flame Metric

10Premixed:

Counterflow:

Agent Effectiveness:

=>can vary with flame type.DMMP

0

0.2

0.4

0.6

0.8

1

0 0.01 0.02 0.03 0.04 0.05

Normilized Bu

rning Ve

locity

DMMP Volume Fraction

CH4

C2H4

CH2O

0

0.2

0.4

0.6

0.8

1

0 0.01 0.02 0.03 0.04 0.05

Normalized

Burning

Velocity

Br2 Volume Fraction

CH4

CH2O

C2H4

Fuel Type Can Influence Inhibitor Effectiveness In premixed flame,

Br2 affected more than DMMP by fuel type (for CH4, C2H4, CH2O)

This image cannot currently be displayed.

Flame Temperature Affects Inhibition Efficiency

=> Cooler flames are inhibited more (for either Fe or DMMP).

0.0

0.2

0.4

0.6

0.8

1.0

0 100 200 300 400 500Fe(CO) 5 Mole Fraction (ppm)

Nor

mal

ized

Bur

ning

Vel

ocity

0.20

= 1.0X = 0.24O2,ox

0.21

0

0.2

0.4

0.6

0.8

1

0 0.003 0.006 0.009 0.012 0.015

Normalized

Burning

Velocity

OP(OH)3 Volume Fraction

1 0.53

1.33

=1.14

0

0.2

0.4

0.6

0.8

1

0 0.005 0.01 0.015 0.02 0.025 0.03

Normalized

Burning

Velocity

Br2 Volume Fraction

0.53

1.33

1.141

Stoichiometry Affects Inhibitor Efficiency

=>lean flame inhibited much more for Br2 than for H3PO4.

HOPO2 HOPO

PO2PO3

OP(OH)2

+O

+OH,O+OH+H

+O2

+H+H+H +H

+O

Lean Stoic.

P catalytic inhibition cycle

Stoichiometry Affects Inhibitor Efficiency

=>can be explained by thermodynamics

1. Characteristics of flame system affect efficiency of chemical inhibition (via flicker, heat losses/stand-off, stabilization mechanisms, mixing conditions, etc.), BUT

2. Inhibitors can influence the physical properties of the flame itself (where and how the flame stabilizes, temperature, flow field, etc.) and hence the inhibition.

3. Most of the physical properties of the flames over polymers which can affect chemical inhibition have not been studied (even more so if a condensed-phase FR is also working).

Bottom Line:

1. Have excellent tools for understanding gas-phase action of fire retardants.

2. Use detailed modeling to understand:a. Why DMMP did not work well in cup burner.b. Why antimony / bromine systems does work well

- Sb? Br? Why no condensation?

3. Study flames with polymers with FRs.

4. Use numerical code to model burning cylinder (simulating UL-94)

4. Systems of interest?

Future Work:

Numerical simulations to understand UL-94

Porous cylinder flames Resemble solid burning (e.g., UL 94) Perform time-dependent axisymmetric

computation with full chemistry/transport for:

DMMP added to fuel Br2 added to fuel

CH4 in air 15% CO2

PMMA in air 18% CO2

Questions?

influence of flame characteristics on gas- phase fire

Documents