Shock Tube Studies of Liquid Fuel Combustion Kinetics

Fuel Summit at USC, September 15-17, 2009

R. K. Hanson, D. F. Davidson, S. S. Vasu, D. Haylett, K.-Y. LamDepartment of Mechanical Engineering

f

, p ,

Stanford University

Objective:Improve knowledge of combustion kinetics of vaporized liquid fuels relevant to the Army – through use of advanced h k t b d ti l di tishock tube and optical diagnostics

Work supported by the ARO Core programDr Ralph Anthenien: Contract MonitorDr. Ralph Anthenien: Contract Monitor

Outline of TopicsOu o op s

1. Development of second-generation aerosol shock tube p gand its application to n-dodecane and diesel ignition

2. Ignition measurements of cyclic jet fuel surrogate components: toluene and methylcyclohexane (MCH)components: toluene and methylcyclohexane (MCH)

3. Extension of the operating regime and capabilities of conventional shock tubes

4. Successful resolution of apparent discrepancy in low-T propane ignition delay times

2

Stanford Shock Tube & Laser Facilities

Transmitted Beam Detector

PressureGC

D2Lamp

DriverSection

Shock TubeDriven Section

PressurePZTPMT

Excimer photolysisVS VRS

P5T5

UV/Vis/IREmissionDetectors

sideKinetic

SpectrographARASMRAS

Shock Tubes (4)

Large Diameter Tubes (15 cm and 14 cm)

High Pressure Tube (5 cm) heatable to 150C

Ring Dye Lasers

Diode Lasers(Near IR & Mid-IR)

g essu e ube (5 c ) ea ab e o 50

Aerosol Tube (11 cm)

Optical Diagnostics

Absorption (VUV, UV, Vis, Near‐IR, Mid‐IR)

Kinetic Spectrograph (200‐300nm)

Ti:Sapphire Laser(Deep UV)

(UV & Vis)

CO2 Lasers(9.8-10.8 m)

Kinetic Spectrograph (200‐300nm)

Emission (UV, Vis, IR)

ARAS (H/O/C/N‐atom detection)

Supporting Equipment

G Ch t h f i it F l A l i

He-Ne Laser(3.39 m)

Gas Chromatography for in situ Fuel Analysis

Excimer Photolysis for Radical Production

Topic 1: Development of Second-GenerationAerosol Shock Tube (AST 2)

First-generation aerosol shock tube (AST 1)

Aerosol Shock Tube (AST 2)

Turbulence-based filling method – some limitations on fuel loading range, uniformity and reproducibility

Second-generation aerosol shock tube (AST 2)Second generation aerosol shock tube (AST 2) New plug-flow-based aerosol delivery system allows partial ST filling Improved spatial uniformity of the aerosol by pre-mixing test mixture Greater range of aerosol (i.e. fuel) loadings

Recent work Fuel 1: N-Dodecane Fuel 2: Diesel (DF-2)

4

Second-Generation Aerosol Shock Tube:Experimental SchematicExperimental Schematic

(22 liters)

Dual extinction yields vapor concentration & temperature

New aerosol loading scheme uses two large ST gate valves, mixing plenum and dump tank; plug flow loading

p

Laser diagnostics (UV to MIR) provide fuel loading, uniformity measure, and species concentration time-histories

5

Fuel: N-DodecaneIgnition Delay Time Measurementsg y

C12H26/21% O2/Ar, =0.5, 5.5 atm 10

= 0.3 = 0.5 = 1 0m

s]

1

= 1.0 = 1.5

elay

Tim

e [

mIg

nitio

n D

e

5.5 atmDodecane/21%O2

0.7 0.8 0.9 1.0 1.1 1.2 1.30.1

1000/T [1/K]

2

Quiescent shock tube behavior, similar to conventional gas-phase ST Low-scatter ignition delay time data Wide range of fuel loading ( = 0 3 to 1 5 in equivalent air) Wide range of fuel loading ( = 0.3 to 1.5 in equivalent air)

6

Fuel: N-DodecaneComparison with JetSurF Modelp

JetSurF model predictions are l i “ l ”

10000

JetSurFonly in “general agreement” with high-P, low-T ignition delay time data

1000

AST 2

Tim

e [m

]

Further work needed to expand range of data (P,T,)

1000

nitio

n D

elay

Dodecane/21%O2

0.8 0.9 1.0100

Ign

1000/T [1/K]

5.5 atm, =0.3

1000/T [1/K]

AST 2 can provide unique vapor-phase data for low-

7

p q p pvapor-pressure fuels in support of mechanism development and validation

Fuel: Diesel (DF-2)Ignition Delay Time MeasurementsIgnition Delay Time Measurements

New diesel ignition data (domestic DF-2, C.I.=55) in excellent agreement with previous Stanford study

Low scatter measurements allow better quantification of ignition time variation with fuel

i ti ( itivariation (e.g. composition, Cetane Index)

Domestic DF-2C.I. = 55

Next Step: - acquire data over wider range of conditions (T,P, mixture, )

8

Next Step: acquire data over wider range of conditions (T,P, mixture, )- investigate jet fuel surrogates (including bio-derived fuels)

Fuel: Diesel (DF-2)First AST 2 Application of Multi-Species Laser Absorption Strategypp p p gy

Multi-species laser absorption measurements

Diesel Ignition

absorption measurements provide new chemical kinetic targets for diesel surrogate comparison

Multi-Wavelength Diagnostics: 650 nm: droplet 650 nm: droplet

extinction 3.39 m: fuel

concentration10 6 h lReflected Shock Conditions: 1092K, 7.6 atm,

f f

10.6 m: ethylene Reflected Shock Conditions: 1092K, 7.6 atm,0.19% Diesel (DF-2), 21% O2

Multi-species data provides time histories for fuel decomposition and stable intermediates (C2H4)

9

Topic 2: Cyclic Jet Fuel Surrogate ComponentsTopic 2: Cyclic Jet Fuel Surrogate Components

High-quality ignition and species data needed for primary surrogate jet fuel components: paraffins (n-dodecane) cyclo-paraffins (methylcyclohexane/MCH) cyclo paraffins (methylcyclohexane/MCH) aromatics (toluene)

Critical need for low-intermediate T data Current targets:

MCH: 60 atm 700 1100 KAromatics

Paraffins MCH: 60 atm, 700-1100 K Toluene: need to resolve ST/RCM

ignition time differences near 1000K

Paraffins

Cyclo-paraffins

10JP-8 Composition

Ignition Measurements of Jet Fuel Surrogate Components: TolueneComponents: Toluene

Development/refinement of surrogate jet fuel mechanisms i li bl i iti d t f t l t i l t

Th P bl R ST d1111K

1000K1250K

requires reliable ignition data for toluene at engine-relevant conditions

The Problem: Recent ST and RCM studies of toluene ignition show differing trends at intermediate temperatures

10000

[s]

Mittal & Sung (RCM)EA= 28.4 kcal/moleToluene/Air

50 atmintermediate temperatures (900-1200 K) and high pressures (near 50 atm)

1000

Davidson et al.EA= 9.6 kcal/moleD

elay

Tim

e = 1

The Need: Establish reliable data base for ign, especially at T< 1050K

100(Shock Tube) SU

Shen et al. (Shock Tube) RPIEA= 13.6 kcal/mole

Igni

tion

D

11

0.75 0.80 0.85 0.90 0.95 1.00 1.05A

1000/T, [1/K]

New Shock Tube Ignition Delay TimeMeasurements: TolueneMeasurements: Toluene

To address this discrepancy we have measured ignition delay times for toluene/air mixtures = 1 0 50 atm 966-1211 Kfor toluene/air mixtures = 1.0, 50 atm, 966 1211 K

New measurements in good agreement with the previous 1111K

1250K 1000Kg pStanford study (Davidson 2005)

Some variation in pressure time-1000

Toluene/Air50 atm= 1

?

phistories with different fuel loading protocol, but no difference in ignition times

y Ti

me

[s]

= 1

New data also consistent with work by Shen (except for their two lowest T pts) 100

Davidson et al. Current Study Shen et al. Best Fit

gniti

on D

elay

12

lowest T pts)0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10

100I

1000/T [1/K]

Toluene Ignition Delay Time Measurements:Comparison with ModelsComparison with Models

General agreement with Andrae et al. (2008) model Shorter experimental times than predicted by Sivaramakrishnan (2005) but Shorter experimental times than predicted by Sivaramakrishnan (2005), but

mechanism employs older rates for key reactions Stanford Modified Model

Stanford Modified Modeluses updated rates for:uses updated rates for:

H+O2=O+OH (GRI)Tol+H=C6H5CH3 (Baulch)Tol+H=C6H5CH2+H2 (Stanford)T l H C H CH (B l h)

10000 910K1110K1000K

s]

1250K

Si k i h (2005)

Andrae (2008)Toluene/Air50 atm, = 1

Tol+H=C6H6+CH3 (Baulch)Tol+O2=C6H5CH2+HO2 (Stanford)Tol+OH=C6H5CH2+H2O (Seta)Benzyl+C6H5CHO= Tol+C6H5CO (Bounaceur)

1000

ay T

ime,

[ Sivaramakrishnan (2005) ?

Good agreement withStanford Modified Model

Next step: acquire data at lower

100Stanford Modified Model (2009)

Igni

tion

Del

a

All ExperimentsStanford & Shen et al.

13

p qtemperatures (

Ignition Measurements of Jet Fuel Surrogate Components: Methylcyclohexane (MCH)Components: Methylcyclohexane (MCH)

Cyclo-alkane (naphthenes) are important chemical class present in jet fuel; ignition time and species data needed for this

Current study provides extended 769K1000K 625K1667K

in jet fuel; ignition time and species data needed for this component

y ptemperature and pressure range of data (P=1.5, 20 and 45 atm)

20 atm shock tube10

100

15atme [m

s]

Pitz et al. (2007) RCM

10atm

20 atm shock tube measurements appear to extrapolate smoothly to RCMmeasurements

1 45atm

20atm

on D

elay

Tim

e1.5atm

Next Step:Testing and refinement ofmulti-component surrogates 0 6 0 8 1 0 1 2 1 4 1 6

0.01

0.1Igni

ti

Current Studies, ST

14

multi-component surrogatesand models

0.6 0.8 1.0 1.2 1.4 1.61000/T [1/K]

Cyclic Jet Fuel Surrogates Components: SummaryCyclic Jet Fuel Surrogates Components: Summary

High-quality ignition and some species time-history data now available for surrogate jet fuel components: paraffins (n-dodecane) cyclo-paraffins (methylcyclohexane/MCH) cyclo paraffins (methylcyclohexane/MCH) aromatics (toluene)

Next Steps: Low-T data needed for toluene Additional species time-history

AromaticsParaffins Additional species time history

data needed for all components(especially at low T)

Paraffins

Cyclo-paraffins

15JP-8 Composition

Topic 3: Extension of Conventional Shock Tube Capabilities

Opportunities/Needs to extend/improve shock tube pp poperating regime

St f d t t Stanford strategy:A. Extend test time using tailored gases and longer driverB. Use driver inserts to correct for dP/dt during RS experimentsB. Use driver inserts to correct for dP/dt during RS experimentsC. Measure T directly using WMS CO2 laser absorption D. Use CHEMSHOCK gasdynamic model to correct for changing test

conditionsconditions

16

Strategy A: Use Driver Extension & Tailored Mixturesto Extend Test Time

Motivation Facility Test Time

Critical need to overlap ST and RCM test times to resolve ignmodeling differences

H hi d 40 !

100

1000

[ms]

2500K 1000K 625K

RCM

10atm3atm

Have achieved > 40 ms!

Shock Tube with Driver Extension 1

10

Igni

tion

Tim

e

UnmodifiedStanford ST

ModifiedStanford ST

0.01

0.1Te

st T

ime,

In-Dodecane/Air =1

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.81E-3

1000/T [1/K]

17(Facility modifications funded by ARO)

Strategy B: Use Driver Inserts to Improve Uniformity of Reflected Shock Test ConditionsReflected Shock Test Conditions

PDriven SectionVS

P5T5

6The Problem: Boundary layers and

attenuation of VS cause slow variations in P5, T5 with time

5

6

dP/dt = 2.2%/ms

)

912K, 4 atm, pure Ar, w/o insert

5, 5

The Solution: (1) Driver inserts: compensates

3

4

897K, 4 atm, pure Ar, w insertdP/dt = 0%/ms

essu

re (a

tm)

for non-ideal variation in pressure (dP/dt)(2) Confirm uniformity with new sensiti e T diagnostic

1

2Pre

L(Driver) = 3.3 m

18

sensitive T diagnostic0 5 10 15

0

Time (ms)

( )

Strategy C: Confirm Uniformity usingNew High-Sensitivity T Diagnosticg y g

Detail of T-HistoryReflected Shock T, P Histories( t il i +i t h t d i )

900

1200

1.8

2.4

ture

[K]

10 ms20

30

40 (uses tailoring+insert, short driver)

600 1.2

[atm

]

Tem

pera

t

-10

0

10 5 K

T

[K]

0 K

0 5 100

300

0.0

0.6952 K, 1.20 atm2% CO2 / Ar Pr

essu

re [

0 5 10-40

-30

-20 952K, 1.2 atm2% CO2/Ar

Direct measurement of T using wavelength modulation spectroscopy (WMS) of CO2 at 2.7m provides validation of highly uniform temperature over long test

Time [ms]0 5 10

Time [ms]

2 p g y p gtimes (T < 3K)

(T-diagnostic developed under AFOSR support)19

Strategy D: Application of new reacting flow model:CHEMSHOCKCHEMSHOCK

Testing/validation of new high-fidelity reaction mechanisms with shock tube data demands both high-quality data and moreshock tube data demands both high quality data and more accurate shock tube reactive gasdynamic models

Improved model CHEMSHOCK Improved model, CHEMSHOCK Uses CHEMKIN plus measured pressure instead of using traditional

assumption of constant U,V

Enable simulation of the temporal evolution of T and species concentrations Enable simulation of the temporal evolution of T and species concentrations behind reflected shock waves with chemical reaction and energy release

Demonstrated successfully with low T ignition of hydrogen and propane

Work in progress to extend CHEMSHOCK to allow simulation of the temporal and spatial variation of non-uniform (1-D) reflected-shock flowfields, including facility effects, as well as, chemical energy release

20

Topic 4: Resolution of Low-T PropaneIgnition Delay Time Discrepancyg y p y

Low T propane ignition i d f ST d RCM

1000 833K

Petersen (2007)

1429K 1000K 714K

time data for ST and RCMboth inconsistent with const. U,V

10

100

[ms]

Herzler (2004) (Ar) Herzler (2004) Gallagher & Curran (2007) (RCM) Const.U,V

Herzler (2004) and Petersen (2007) show strong “NTC-type”

1

10

ition

Tim

e Model

strong NTC type behavior below 1000 K

Stanford strategy: 0.1 2.1% C3H8 / O2 / N2

= 0.5, 30 atm

Ign

gymeasure propane ignition times at 6 & 60 atm using new ST techniques and

0.6 0.8 1.0 1.2 1.4 1.60.01

1000/T [1/K]interpret using CHEMSHOCK model

21

Use of Driver Inserts to Maintain Uniform P and T during Reflected Shock Wave Experiments: 6 atm

151250K 1000K1111K

50

Propane Ignition: Low P Comparison with Model

10

[atm

]

1034KdP/dt = 2.0%

10

dP/dt=0%/ms

dP/dt=1-2%/ms

e [m

s]

Curran et al. (2008)Model

51021KdP/dt = 0.3%

Pre

ssur

e

0.8% Propane/O2/Ar=0 5 7 atm

Igni

tion

Tim

e

0.8% Propane/ O2/ Ar = 0 5 6 atm

0 5 10 15 20 25

0

Time [ms]

=0.5, 7 atm

0.8 0.9 1.0 1.11

1000/T [1/K]

0.5, 6 atm

Experiments performed with and without driver inserts Without driver insert dP/dt = 1-2%/ms (requires CHEMSHOCK model) With driver inserts dP/dt = ~0%/ms (can use traditional constant U,V model)

Primary conclusion: Same mechanism (Curran et al 2008) successfully Primary conclusion: Same mechanism (Curran et al. 2008) successfully simulates all low P data! No discrepancy between detailed propane mechanisms and low T ST data.

22

Future Worku u o

Ignition chemistry at low temperatures (NTC):Species time-histories and ignition delay times Fuels: Toluene, n-Dodecane

Species Monitored: OH, C2H4, H2O, CO2, fuel, benzyl

Application of AST 2:Practical, Synthetic and Bio-derived Jet Fuels

Fuels: S 8 Large (bio oil derived) Methyl Esters Fuels: S-8, Large (bio-oil derived) Methyl Esters

23

Acknowledgementsg

Visit Stanford website http://hanson.stanford.edu/ forF d t l Ki ti D t b U i Sh k T b V l 1 I iti D l Ti V l 2 C t tiFundamental Kinetics Database Using Shock Tubes; Vol. 1: Ignition Delay Times, Vol. 2: Concentration Time Histories, Vol. 3: Reaction Rate Measurements

Work supported by:ARO Core Program for Basic and Applied Scientific Research (Contract Monitor: Dr Ralph Anthenien)ARO Core Program for Basic and Applied Scientific Research (Contract Monitor: Dr. Ralph Anthenien)

Our Shock Tube and Laser Jockies

Dan Haylett Subith Vasu Brian Lam24