Correlation between molecular structure and globalCorrelation between molecular structure and global combustion properties for surrogate fuel modeling:

Extinction Limit, Radical Index, and Flame Speed

Sang Hee Won, Wenting Sun and Yiguang Ju*

Department of Mechanical and Aerospace EngineeringPrinceton University

Sept.20-24, 2010

Introduction: Surrogate Fuel Model DevelopmentS th ti f l h t b d b bl di ith j t f l• Synthetic fuels have to be used by blending with jet fuel: – normal, branched, and cyclo alkanes, and aromatics

• MURI team selected surrogate jet fuel components– n-decane, methyl-cyclohexane, n-propyl- and trimethyl- benzenes,n decane, methyl cyclohexane, n propyl and trimethyl benzenes,

toluene…

• Goals: – Identify an aromatic fuel candidate for surrogate component

Obtain fundamental flame data– Obtain fundamental flame data– Identify the correlation between fuels with different molecular

structures and global burning properties for surrogate fuel model construction

Previous work and findings

• Strong kinetic coupling between n-decane and toluene.

• Diffusion flame extinction of n-decane/toluene is governed by the peak OH radical concentration.

• Fuel chemistry and transport affect diffusion flame extinction limit.

• A multi-timescale (MTS) and multi-generation path flux analysis model reduction method were developed.

18001 50E 04

Example: Complexity in Flame Chemistry:Kinetic coupling between n-decane and toluene in flames

1500

1800

1.00E-04

1.50E-04

s]60 % n-decane + 40 % toluene

near extinction, Xf = 0.1, a = 176 1/s H Diffusion loss

1200

1500

5.00E-05 Temle/c

m3 s

C6H5CH2+ H = C6H5CH3

900

0.00E+00

mperatte

[mol

overall decomposition

600-1 00E-04

-5.00E-05

ure [K]tio

n ra

t overall decomposition of n-decane

overall decomposition

300-1.50E-04

-1.00E-04 ]Re

act

C10H22=C2H5+C8H17-1C10H22=nC3H7+C7H15-1C10H22=CH3+C9H19-1C6H5CH3+H<=>C6H5CH2+H2

pof toluene

0-2.00E-040.85 0.9 0.95 1 1.05 1.1

C6H5CH3+H<=>C6H5CH2+H2C6H5CH3+OH<=>C6H5CH2+H2OC6H5CH3+H<=>A1+CH3

0.1 mm

Axial coordinate [cm]Won, Sun, Dooley, Dryer, and Ju, CF 2010

The questions to answer:• Will n-propyl- and trimethyl- benzenes with same H/C &• Will n-propyl- and trimethyl- benzenes with same H/C &

molecule weight and similar “Tad” have different extinction limits?

• Will the MURI generic rule for surrogate fuel model really work for both flames and ignition?g

• How do we create a correlation between global flame properties and fuel chemical and transport properties?properties and fuel chemical and transport properties?

• What are the uncertainties in experimental measurements ofWhat are the uncertainties in experimental measurements of global flame properties?

• How does the PFA model reduction approach work for surrogate fuel mixtures?

1. Experimental Studies of Molecule Structure of Aromatics on Diffusion Flame Extinction Limits

• Four aromatics– Toluene, – n-propylbenzene, 124- and 135- trimethylbenzenesn propylbenzene, 124 and 135 trimethylbenzenes

• Same H/C ratio, molecular weight (C9H12), flame temperature

nPB 135TMBnPB124TMB

135TMB

• Experimental setup: extinction limitsp p– Using FTIR measurements to check thermal decomposition and

concentration variation: Concentration fluctuation < 1%OH radical meas rement Q1(6) e citation 282 93– OH radical measurement: Q1(6) excitation ~ 282.93 nm

Heater & insulation

Nitrogen

Heater & insulationHeater

Fuel

T1 T2

To burner

Schematics of evaporation system

Experimental Measurements of OH Concentration b i l i d d flby using laser induced fluorescence

• Nd:YAG Laser : Quanta-Ray (532 nm), Cobra-stretch dye laser• ICCD Camera : PIMAX-Gen II (Princeton Instrument)• PLIF: Q1(6) transition (282.93 nm), linear regime, beam height 80 mm

Quenching Soot PAH correctionsQuenching, Soot, PAH corrections

PAH LIFLII (soot)

OH LIF

Q1(6) line

Stagnation plane

d t d bt ti

g p

Schematic photo for Rayleigh scattering and PLIF

detuned subtraction

nPB diffusion flames

Extinction Limit Comparison• Extinction limits vary significantly according to fuel structureExtinction limits vary significantly according to fuel structure

– nPB > toluene > 124TMB > 135TMB– Kinetic effects from the specific fuel chemistry!

400toluene_exp.

300a E[1

/s] 124TMB_exp

135TMB_expnPB_exp

n-propyl benzene

t l

200rain

rate

toluene

124 trimethyl benzene

100ctio

n st

r

135 trimethyl benzene

100

Extin

c

00.05 0.1 0.15 0.2 0.25

Fuel mole fraction Xf

Measurements and Prediction Comparison• Poor agreement with TMBs why? Transport or• Poor agreement with TMBs, why? Transport or

kinetics

400toluene124TMB

300

e a E

[1/s

] 124TMB135TMBnPBtoluene (Princeton)124TMB (Aachen)

nPB

toluene

200rain

rate 124TMB (Aachen)

nPB (Princeton)

100nctio

n st

135TMB

0

Extin

124TMB

00.05 0.1 0.15 0.2 0.25

Fuel mole fraction Xf

OH concentrations and comparisons3E-08

on toluene

nPB

Xf = 0.15

2E-08

cent

ratio

cm3 ]

124TMB

1E 08OH

conc

[mol

e/c

toluene (computation)

135TMB2E-08

3E-08

e/cm

3 ]

a = 140 1/s

1E-08

Max

. O toluene (computation)nPB (computation)124TMB (computation)toluene (LIF)nPB (LIF) 0

1E-08

OH

[mol

e

fuelside

oxidizerside

00 100 200 300 400

( )124TMB (LIF)135TMB (LIF)

00.3 0.4 0.5 0.6 0.7

Axial coordinate [cm]

[OH] proportional to the e tinction strain rates

Strain rate a [1/s]

• [OH]max proportional to the extinction strain rates• [OH]max : nPB > toluene > 124TMB > 135TMB

2. Surrogate Model Flame Validation on Jet POSF 4658POSF 4658

•First generation surrogate model (3 components)

Surrogate Fuel: Mole Fraction DCN H/C MW / g mol-1 TSIJet‐A POSF 4658 47.1 1.957 142.01 21.4

n-decane iso-octane Toluene0.4267 0.3302 0.2431 47.1 2.01 120.7 14.1

Mechanism compilation: n-decane/iso-octane/toluenen decane => Westbrook et al (LLNL) 2008n-decane => Westbrook et al. (LLNL) 2008.iso-octane => Mehl et al. (LLNL revision of Curran et al.) 2010.Toluene => Princeton, (Metcalfe, Dooley and Dryer) 2010.C0-C4 assembled and tested at Princeton0 4

•Second generation surrogate model (4 components)•n-dodecane/iso-octane/ propyl benzene / 1,3,5 trimethyl benzene

12

Validation: Extinction of Diffusion FlamesPOSF 4658 fi t ti tPOSF 4658 vs. first generation surrogate

Fuel mass fraction, Yf

5000.2 0.3 0.4 0.5 0.6

Princeton JetA surrogateJetA1/

s]

,f

300

400ra

te, a

E [1

200

on s

trai

n

100

Extin

ctio

00 0.03 0.06 0.09 0.12 0.15

Fuel mole fraction, Xf

Validation: Extinction of Diffusion FlamesPOSF 4658 d ti tPOSF 4658 vs. second generation surrogate

400

]

300

e a E

[1/s

]

Why?

200rain

rate

JETAPOSF 4658

100nctio

n st

JETA POSF 4658

3 comp. Surrogate

4 comp. Surrogate100

Extin

co p Su ogate

00.2 0.3 0.4 0.5

Fuel mass fraction Yf

3. Correlations for extinction limit: radical Index

500

/s]

n-decanen-nonanen-heptane

n-alkanesTf = 500 K and To = 300 K

400

te a

E[1

/ n heptaneJETA POSF 4658Princeton Surrogateiso-octanenPB

300

rain

rat nPB

toluene124TMB135TMB

aromatics

200

ctio

n st

r

100

Extin

c

00 0.05 0.1 0.15 0.2

F l l f ti XFuel mole fraction Xf

The role of radicals?2 2E 08

2E-08

2.2E-08/c

m3 ] toluene

nPB124TMB

1 8E-08

2E-08

[mol

e/ 124TMB

1.6E-08

1.8E-08

n lim

it

1.4E-08

1.6E 08

tinct

ion

1.2E-08at e

xt

1E-08

[OH]

Possibility to define a radical index for scaling?

0 100 200 300 400

Extinction strain rate aE [1/s]

4. Uncertainty in flame speed measurements: Spherical flames vs. Counterflow flames

no flame

Q Ub

r

1.2

1.4

1.6

m/s

)

x=3.2mm +/- 3%

with flame3035

m/s

)

0 4

0.6

0.8

1

xial

vel

ocity

(m10152025

ame

spee

d (c

m

suo

ubfu VS ρρ /=

0

0.2

0.4

-8 -6 -4 -2 0 2 4 6 8

Radius (mm)

Ax

r

05

0 100 200 300 400 500 600 700Stretch (1/s)

Fla

Radius (mm)

Good experiments?

4. Uncertainty in flame speed measurements: Spherical flames: Effect of flow compression, ignition, and stretch

y,a*

U/S

L

0.4

0.6

0.8

Q Vb

aliz

edflo

wve

loci

ty

0 2

0

0.2

N li d ti l di t /R

Nor

ma

0 0.25 0.5 0.75 1-0.6

-0.4

-0.2

Nomalized spatial coordinate, r/R0

d,s u

(cm

/s)

220

240 0.6Rw0.1Rw0.2Rw0.3Rw0.4Rw

Hydrogen-air, 1 atm, φ = 3.0 Ignition energy

flam

esp

eed

180

200

220

Cal

cula

ted

0 1000 2000 3000

160

180

Flow-correctedUncorrected

)/(*/)( ,, bubuubbfu VVS ρρρρ−=Stretch rate, κ (1/s)

50

Uncertainty in flame speed measurements: n-heptane/airSpherical flame: flame speed extraction

45

50

s)L

40

45

, SL (c

m/s

S 39 4 /

Unstreched 1D flame speed (PREMIX)

35

40

me

spee

d, Spherical flame SL0=39.4 cm/s

Rf=2.5 cm

30

35

ched

flam

Ignition energy driven regionRf=1.5 cm

25

30

Stre

t

Linear regionRf 1.5 cm

0 200 400 600 800 1000 120025

Stretch rate, s-1

Spherical flame modeling: Multi-generation PFA + MTS + A-SURF1D(Sun, Gou, Chen, and Ju, CF, 2010)

engine/princeton.edu/downloads

Uncertainty in flame speed measurements: n-heptane/airCounterflow flames: flame speed definition and burner separation distance

50SL, Umin

ty (K

)440Plug flow limit

45

ow v

eloc

it

L (cm

/s) L

Spherical flame POTN L=4 400

420

40

nim

um fl

o

spee

d, S

POTN L=4 PLUG L=4 cm PLUG L=1 cmPLUG L=0 6 cm

SL0=39.4 cm/s 380

35

S ure

at m

in

ed fl

ame PLUG L=0.6 cm

SL, w01340

360

30SL, wmax

Tem

pera

t

Stre

tche

320

3 0

0 200 400 600 800 1000 120025

T

Stretch rate s-1

300

Stretch rate, s

5. Model reduction for surrogate fuel mixtures using multi-generation PFA method

Challenges:•Multiple fuel components and surrogate targets: 3-5 fuelsMultiple fuel components and surrogate targets: 3-5 fuels•Unstable flame regimes for large molecule fuels

AA

M1 M2 M3 Mn…

B C

Objectives:Examine the potential of the multi-generation Path fluxExamine the potential of the multi generation Path flux analysis (PFA) method for jet fuel surrogates

Chem-RC (PFA) download site:Chem-RC (PFA) download site: http://engine.princeton.edu

Target models: CSE Kerosene and MURI Surrogates

•(n-Decane, iso-Octane, n-Propyl-Benzene) > 1000 species•(MURI first generation surrogate model) >1000 species

Validation of Kerosene surrogate fuels

10n-Propyl-Benzene /air, Φ=1

1

10fuel mixture/air, Φ=1

n Decane 39 03%

0.1

10

P=0.3 atm

P=1 atm

y τ ig

(sec

)

P 5 t

0.1

10P=0.3 atm

P=1 atm

ay τ ig

(sec

)

P 5 t

n-Decane - 39.03%iso-Octane - 24.92%n-Propyl-Benzene - 36.05%

1E-3

Igni

tion

dela

y P=5 atm

1E-3

Igni

tion

dela P=5 atm

0.4 0.6 0.8 1.0 1.2 1.4 1.61E-7

1E-5I

Detailed mechanism Reduced (547 species) Reduced (316 species)

0.4 0.6 0.8 1.0 1.2 1.4 1.61E-7

1E-5I

Detailed mechanism Reduced (547 species) Reduced (316 species)

1000/T (1/K)0.4 0.6 0.8 1.0 1.2 1.4 1.6

1000/T (1/K)

Validation of MURI first generation surrogate

70s] Tu = 400 K

60

S L[c

m/s u

50

peed

S

40

flam

e s

30

min

ar f

Experiments [JETA POSF4658]

200 6 0 8 1 1 2 1 4 1 6

Lam

Model_154 species

0.6 0.8 1 1.2 1.4 1.6

Equivalence ratio φ

Conclusion• Trimethyl benzenes and n-propyl-benzene have the same H/C ratio

d l l i ht b t diff t ti ti li itand molecular weight but different extinction limits.

• Extinction limits are proportional to the maximum OH concentrations, and strongly affected by transport and fuel heating value.and strongly affected by transport and fuel heating value.

• A linear correlation between extinction limit and radical index, heating value, and molecular weight was proposed and validated for fuels

ith diff t l l t twith different molecular structures.

• Measurements of intermediate species and radicals in flames are important to provide additional constraints for kinetic mechanisms.important to provide additional constraints for kinetic mechanisms.

• The first generation jet surrogate model works well in reproducing extinction limit of real fuel in molar correlation, and the second

ti t k ll f b th l d l tigeneration surrogate works well for both molar and mass correlations.

• Uncertainty in flame property measurements remains to be a challenging issue and needs to be addressed in a unified way.challenging issue and needs to be addressed in a unified way.

• The multi-generation path flux method for model reduction were tested for a kerosene surrogate with multi-component fuel mixtures, and integrated to multi timescale method with adaptive mechanismsand integrated to multi-timescale method with adaptive mechanisms.

Thanks Dr. Tishkoff for the MURI research grantg

A flame evolution theoryAssumptions:

Qig ?Assumptions:

Constant densityOne-step chemistryCenter energy deposition

Theory:

⎤⎡22

Center energy depositionChen and Ju, 2007

⎥⎥⎦

⎤

⎢⎢⎣

⎡

−+

−==⋅−⋅

∫∫∞

−−

−−−−

∞−−

−−

f

f

ULe

ULeRUR

U

UR

f TTZ

de

eRLe

eRde

eRT)(

expσσ

ττττ ττ 11

21Q

2

22

ig2

2

∫∫RR

dede ττττ

2 12122 Z

Large flame

Nonlinear Model: )()()R2ln()( 112112

22

−−−⋅=++LeRLeR

ZUR

U

Small velocity derivation

Linear Model: KaMaSS uu ⋅−= 1/ 0

Small velocity derivationNot included in Sivashinsky, 1979

Sivashinsky, 1979