Lawrence Livermore National Laboratory

Lawrence Livermore National Laboratory

Chemical Kinetic Research on HCCI & Diesel Fuels

William J. Pitz (PI), Charles K. Westbrook, Marco Mehl, Mani SarathyLawrence Livermore National Laboratory

May 15, 2012

DOE National Laboratory Advanced Combustion Engine R&D Merit Review and Peer Evaluation

Washington, DCThis presentation does not contain any proprietary, confidential or otherwise restricted information

This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344

Project ID # ACE013

LLNL-PRES-543385

2LLNL-PRES- 543385 2012 DOE Merit Review


Overview

• Project provides fundamental research to support DOE/ industry advanced engine projects

• Project directions and continuation are evaluated annually

Technical Barrier: Increases in engine efficiency and decreases in engine emissions are being inhibited by an inadequate ability to simulate in-cylinder combustion and emission formation processes• Chemical kinetic models for fuels are a

critical part of engine models Targets: Meeting the targets below relies

heavily on predictive engine models for optimization of engine design:• Fuel economy improvement of 25 and 40%

for gasoline/diesel by 2015• Increase heavy duty engine thermal

efficiency to 55% by 2018.• Attain 0.2 g/bhp-h NOx and 0.01 g/bhp-h PM

for heavy duty trucks by 2018

Project funded by DOE/VT:• FY11: 500K• FY12: 640K

Timeline

Budget

Barriers/Targets

• Project Lead: LLNL – W. J. Pitz (PI), C. K. Westbrook, M. Mehl, S. M. Sarathy

• Part of Advanced Engine Combustion (AEC) working group:

• – 15 Industrial partners: auto, engine & energy• – 5 National Labs & 2 Univ. Consortiums• Sandia: Provides HCCI Engine data for validation of

detailed chemical kinetic mechanisms• FACE Working group

Partners



Objectives and relevance to DOE objectives Objectives:

• Develop predictive chemical kinetic models for gasoline, diesel and next generation fuels so that simulations can be used to overcome technical barriers to low temperature combustion in engines and needed gains in engine efficiency and reductions in pollutant emissions

FY12 Objectives:• Develop detailed chemical kinetic models for larger alkyl aromatics• Develop a reduced surrogate mechanism for diesel to be used for

multidimensional CFD simulations • Develop more accurate surrogate kinetics models for gasoline • Develop a functional group method to represent cycloalkanes in diesel

fuel• Validate and improve 2- and 3-methyl alkanes mechanisms with new

data from shock tubes, jet-stirred reactors, counterflow flames, and premixed flames



Chemical kinetic milestones January, 2012

Develop more accurate surrogate kinetics models for gasoline • June, 2012

Develop a functional group method to represent cycloalkanes in diesel fuel• June, 2012

Develop detailed chemical kinetic models for larger alkyl aromatics• September, 2012

Develop a reduced surrogate mechanism for diesel to be used for multidimensional CFD simulations

• September, 2012Validate and improve 2- and 3-methyl alkanes mechanisms with new data from shock tubes, jet-stirred reactors, counterflow flames, and premixed flames



Approach Develop chemical kinetic reaction models for each individual fuel component of

importance for fuel surrogates of gasoline, diesel, and next generation fuels

Combine mechanisms for representative fuel components to provide surrogate models for practical fuels• diesel fuel• gasoline (HCCI and/or SI engines)• Fischer-Tropsch derived fuels• Biodiesel, ethanol and other biofuels

Reduce mechanisms for use in CFD and multizone HCCI codes to improve the capability to simulate in-cylinder combustion and emission formation/destruction processes in engines

Use the resulting models to simulate practical applications in engines, including diesel, HCCI and spark-ignition, as needed

Iteratively improve models as needed for applications



Technical Accomplishment Summary

Development of chemical kinetic model for larger aromatics

Determined effect of branching for alkanes on ignition under engine conditions

Validated approach and mechanism for gasoline surrogate fuels

RCM

Shock tube

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

0.7 1.2

Igni

tion

Del

ay T

ime

(s)

1000/T (1/K)

Octane isomersphi=1, 20 atm, in air

2mhp

nc8

3mhp

0.1

1

10

100

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

Current WorkRCM SimulationGauthier et al. (2004)Constant Volume Simulation

Tota

l Ign

ition

Del

ay (m

s)

1000/TC (K-1)

(b) φ=1.0, PC=40 bar

Developing reduced chemical kinetic model for diesel for engine combustion network (ECN)

0

200

400

600

800

1000

1200

550 600 650 700 750 800 850

Temperature [K]

CO

[ppm

]

Flow Reactor Data8 bar

Ф = 0.23(23% m-xylene /

77% n-dodecane)



Fuel Surrogate Palette for Diesel

n-alkanebranched alkanecycloalkanesaromaticsothers

butylcyclohexanedecalin

hepta-methyl-nonane

n-decyl-benzenealpha-methyl-naphthalene

n-dodecanen-tridecanen-tetradecanen-pentadecanen-hexadecane

tetralin

Need representative component models to fill outthe diesel fuel surrogate palette:

trimethyl-

2,9-dimethyldecane

2-methylpentadecane

3-methyldodecane

Lightly methylated iso-alkanes

Large aromatics



Jet Stirred Reactors

• Idealized chemical reactors with/without simplified transport phenomenon

Non Premixed Flames

Premixed Laminar Flames

Engine Combustion

Combustion Parameters

Temperature

Pressure

Mixture fraction (air-fuel ratio)

Mixing of fuel and air

Need to validate the fuel component and surrogate models

Shock tube

Electric ResistanceHeater

Evaporator

Fuel Inlet

Slide Table

Oxidizer Injector

Optical Access Ports

Sample ProbeWall Heaters

High pressure flow reactors

Rapid Compression Machine

9 LLNL-PRES- 543385 2012 DOE Merit Review!


Modeling of 3-methylheptane Ignition

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

0.7 0.9 1.1 1.3 1.5

Igni

tion

Del

ay T

ime

(s)

1000/T (1/K)

3-methylheptane 20 atm, in air

3mhp phi=0.5 3mhp 3mhp phi=1.0 3mhp 3mhp phi=2 3mhp

Model is generally within a factor of 1.5-2 of the experimental data.

Model slightly faster than shock tube data in the NTC region

3-methylheptane

!=0.5 !=1

!=2

Shock tube

W. Wang, Li, Oehlschlaeger, Healy, Curran, Sarathy, Mehl, Pitz, Westbrook, Proc. Combust. Inst. (2012).



Modeling the effect of branching on ignition

-19.0

21.7

36.8

RON

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

0.7 0.9 1.1 1.3 1.5

Igni

tion

Del

ay T

ime

(s)

1000/T (1/K)

Octane isomersphi=1, 20 atm, in air

2mhp

nc8

3mhp

iso-octane

100

iso-octane

3-methylheptane

n-octane

2-methylheptane

Shock tube

W. Wang, Li, Oehlschlaeger, Healy, Curran, Sarathy, Mehl, Pitz, Westbrook, Proc. Combust. Inst. (2012).



Jet Stirred Reactor (JSR) 3-methylheptane CNRS Orleans, France

•! The zero dimensional perfectly system is ideal for modeling.

•! The products species profiles are dependent on chemical kinetics due to the perfectly mixed homogeneous environment.



0.0E+00

2.0E-04

4.0E-04

6.0E-04

8.0E-04

1.0E-03

1.2E-03

1.4E-03

500 600 700 800 900 1000 1100 1200

Mol

e Fr

actio

n

Temperature (K)

JSR 3-Methylheptane Results

Good prediction of low T and high T reactivity

Fuel

C2H4

H2

! = 1 10 atm

Dagaut et al. CNRS (2010)

Experiments performed at CNRS, Orleans (F. Karzenty, C. Togbe G. Dayma, P. Dagaut)

Lines: LLNL model

Curves: CNRS experiments

" = 0.7 sec



Model simulations and experiments show decrease in laminar flame speeds with branching for octane isomers

#!Important to properly predict flame speeds for gas turbine applications

#!Location of single methyl branch has minimal effect on flame speed

!"#

$%#

$"#

&%#

&"#

'%#

'"#

"%#

""#

%()# %(*# !(%# !($# !('# !()#

!"#$%"&'()"#*'+,**-.'+/0

'12#345'

67/$8")*%2*'9":0'

+,-.#&"&#/.#!+01#2342-,12506#76819:;6<.#1:=2;#7;,526<#

5>:?0+52###$>120@8;@240+52###&>120@8;@240+52###$.">=,120@8;@23+52###

n-octane

Experiments: symbols

Model: lines

2,5-dimethylhexane Experiments performed at USC, Los Angeles

3-methylheptane

2-methylheptane

Lam

inar

flam

e sp

eed

[cm

/s]

Equivalence ratio

(LLNL)

(USC) Ji, Sarathy, Veloo, Westbrook and Egolfopoulos, Combust. Flame (2012)



Developed new procedure to formulate gasoline surrogates

1.0E-04

1.0E-03

1.0E-02

1.0E-01

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5

Igni

tion

Dela

y Tim

es [s

]25 atm

1000K/T

T=825 K

Experimental data: Personal communication and N. Morgan et al. Comb. & Flame

slope

40 different gasoline surrogates were simulated

Slope in NTC region correlates to octane sensitivity = RON - MON

Ignition delay at 825K correlates to octane number

FY2011



A correlation between the octane rating and the autoignition propensity of gasoline surrogate fuels has been identified

y = -0.60x2 + 2.64x + 8.86R² = 0.79

-2

0

2

4

6

8

10

12

14

16

-3 -2 -1 0 1 2 3 4

PRFs

Higher unsaturated content

Slope

Sens

itivi

ty

a) y = 67.95x3 + 422.17x2 + 903.17x + 753.93R² = 0.99

0

20

40

60

80

100

120

140

-3.5 -3 -2.5 -2 -1.5 -1

PRFs

Ignition Delay Times [Log(1/s)] O

I

b)

Sensitivity vs. Slope Octane Index vs. Ignition Delay time @ 825K

Starting from the analysis of 40 different fuel mixtures a general approach to the formulation of the gasoline surrogate has been developed on the basis of the key features of the ignition delay curves

25 atm

FY2011



Surrogate (%Vol)

RD387 Gasoline (%Vol)

i-Alkanes 57n- Alkanes 16Aromatics 23 23

Olefins 4.0 4.2

73

Matched gasoline properties with a surrogateAlkanes

Aromatics

Olefins

FY2011

A/F Ratio 14.6 14.8H/C 1.93 1.95

Octane index 87 87Sensitivity 8.0 7.6

Composition:

Other properties:



Simulations of real gasoline (RD387) experiments using LLNL surrogate model

0.01

0.1

1

10

100

0.6 0.8 1 1.2 1.4 1.6

Current WorkGauthier et al. (2004)RCM SimulationConstant Volume Simulation

Tota

l Ign

ition

Del

ay (m

s)

1000/TC (K-1)

(a) φ=1.0, PC=20 bar

0.1

1

10

100

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

Current WorkRCM SimulationGauthier et al. (2004)Constant Volume Simulation

Tota

l Ign

ition

Del

ay (m

s)

1000/TC (K-1)

(b) φ=1.0, PC=40 bar

RCM experiments: G. Kukkadapu, K. Kumar, C.-J. Sung, Univ. of ConnecticutShock tube experiments: Gauthier, Davidson and Hanson, 2004

RCM

Shock tube

RCM

RCM model

Constant volume model

Curves: LLNL modelSymbols:○ Univ. Conn. experiments□ Stanford experiments

Curves: LLNL modelSymbols:○ Univ. Conn. experiments□ Stanford experiments

RCM

Shock tube RCM model

Constant volume model



Purely experimental comparisons of LLNL and Stanford surrogate formulation with real gasoline: LLNL formulation matches well real gasoline

0

5

10

15

20

25

30

35

40

-20 -15 -10 -5 0 5 10

(a) φ=1, PC=20 bar, TC=758 K

Pres

sure

(bar

)

Time (ms)

LLNL

Gasoline

Stanford A

0

10

20

30

40

50

60

70

80

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

(c) φ=1, PC=40 bar, TC=688 K

Pres

sure

(bar

)Time (ms)

Stanford A

Gasoline

LLNL

RCM experiments:

G. Kukkadapu, K. Kumar, C.-J. Sung, Univ. of Conn.

(RD387)

(RD387)

RCM



Surrogate for larger alkyl aromatics: n-heptane + n-propylbenzene

6 7 8 9 10 11 12 13 14 15 160.01

0.1

1

10

100

1 atm 10 atm 30 atm

Igni

tion

dela

y tim

e / m

s

10,000 K / T

Curves: NUIG-LLNL model

Experiments: NUIG

1 atm

10 atm

30 atmshock tube

Rapid compression machine

with heat loss model

Experimental RCM data: Darcy and Curran, NUIG, 2011

Shock tube

10 atm

ɸ= 0.98

43% n-heptane /57% n-propylbenzene in air

Shock tube study: Darcy, Mehl, Simmie, Wurmel, Metcalfe, Pitz and Curran, Proc. Combust. Inst. (2012)



Larger aromatics: Shock tube ignition of n-butyl benzene at high pressure

n-butylbenzene

Experiments: Tobin,Yasunaga and Curran, NUIG, Ireland (Combust. Flame 2011)

1E-05

0.0001

0.001

0.01

0.6 0.7 0.8 0.9 1.0 1.1

Igni

tion

dela

y tim

e [m

s]

1000 K / T

1 atm10 atm

30 atm

φ=0.5

Curves: NUIG-LLNL model

Symbols: NUIG experiments

Shock tube



Developing reduced model for diesel surrogate for Engine Combustion Network (ECN) (23% m-xylene / 77% n-dodecane)

(by liquid volume)

CO formation is a good index of the reactivity of the mixtureGood predictions on the peak value, but shifted by 40K Reasonable agreement on the flame speed measurements

Flow reactor data: Natelson, Kurman, Johnson, Cernansky, and Miller, Combust. Sci. Tech. (2011)Flame data: Ji, Moheet, Wang, Colket, Wang, Egolfopoulos, Combustion Meeting, 2011

0

200

400

600

800

1000

1200

550 600 650 700 750 800 850

Temperature [K]

CO

[ppm

]

Flow Reactor Data8 bar

Ф = 0.23

LLNL detailed model results:

010203040506070

0.6 0.8 1.0 1.2 1.4 1.6

Flam

e sp

eed

[cm

/s]

φ

1 atmTunburned = 403K

Flame speed comparisons



Mechanisms are available on LLNL website and by email

LLNL-PRES-427539

http://www-pls.llnl.gov/?url=science_and_technology-chemistry-combustion



Collaborations Our major current industry collaboration is via the DOE working groups on HCCI

and diesel engines • All results presented at Advanced Engine Combustion Working group meetings

(Industry, National labs, U. of Wisc., U. of Mich.)• Collaboration with John Dec at Sandia on HCCI engine experiments with

gasoline • Collaboration with Sibendu Som at Argonne on diesel reacting sprays

Second interaction is collaboration with many universities• Prof. Sung’s group, U of Conn. on gasoline surrogates• 2-methyl, 3-methyl and dimethyl alkanes:

− Prof. Oehlschlaeger at RPI, shock tube ignition at engine pressures − Prof. Egolfopoulos at USC flame speed, ignition and extinction− Prof. Seshadri at UC San Diego: flame ignition and extinction− Dr. Dagaut, CNRS, Orleans, France

• Dr. Curran at Nat’l Univ. of Ireland on 2-methyl, 3-methyl heptane, n-propyl benzene and n-butyl benzene in RCM and shock tube

• Prof. Lu, U. of Conn. on mechanism reduction Participation in other working groups with industrial representation

• Fuels for Advanced Combustion Engines (FACE) Working group and AVFL-18 (Surrogate fuels for kinetic modeling)

• Engine combustion network (ECN)



Activities for Next Fiscal Year

Develop detailed chemical kinetic models for: • larger alkyl aromatics

• larger n-alkanes (above C16)(important to get end of distillation curve)

(both coordinated with CRC AVFL-18, “Surrogate fuels for kinetic modeling”)

Modeling of engine combustion with reduced models for diesel surrogate fuels for the Engine Combustion Network

Gasoline surrogate work with Prof. Sung (UConn) for RCM ignition and Egolfopoulos (USC) for flame speeds• Look at the effect of EGR



Summary Approach to research

• Continue development of surrogate fuel mechanisms to improve engine models for HCCI and diesel engines

Technical accomplishments:• We validated our 3-methyl, 2-5-dimethyl alkane, and alkylated

aromatics mechanisms by comparison to experimental data at engine-like pressures and temperatures

Collaborations/Interactions• Collaboration through AEC working group and FACE working

group with industry. Many collaborators from national labs and universities

Plans for Next Fiscal Year:• Larger alkyl aromatics:• Reduced models for engine combustion network (ECN)• Validate gasoline surrogate mechanism with experiments

including EGR

Documents

Lawrence Livermore National Laboratory