Lawrence Livermore National Laboratory · USC, Los Angeles r flame spe 15 20 Lamina 25 3‐methylheptane 2,5‐dimethylhexane Experiments: symbols Lamina Model: lines (LLNL) (USC)

Lawrence Livermore National LaboratoryLawrence Livermore National Laboratory

Fuel surrogate Mechanisms for Conventional and Biofuels

William J Pitz Marco Mehl Charles K WestbrookWilliam J. Pitz, Marco Mehl, Charles K. WestbrookLawrence Livermore National Laboratory

September 19, 2012

MACCCR5th Annual Fuels Research Review

Combustion Research FacilityLivermore, CA

This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344LLNL-PRES-582236

Acknowledgements

Mani Sarathy, now at KAUST in Saudi ArabiaH C N ti l U i it f I l d G l Henry Curran, National University of Ireland, Galway

Support from program managers Gurpreet Singh and Kevin Stork, Office of Vehicle Technologies,Kevin Stork, Office of Vehicle Technologies, Department of Energy

Support from program manager Wade Sisk, Basic E S i D t t f EEnergy Sciences, Department of Energy

Support from program manager Sharon Beerman-Curtin Office of Naval ResearchCurtin, Office of Naval Research

2LLNL-PRES- 582236 MACCCR 2012

Lawrence Livermore National Laboratory

Progress on surrogate fuels and biofuels

CH3CH3

Defining surrogates for gasoline fuels: Detailed chemical kinetic models for surrogates components for diesel fuels:

1-methylnaphthalenen-hexadecane

-CH3C

aromaticstetralin

1-methylnaphthalene

1,2,4-trimethylbenzene

decalin

n-dodecylcyclohexane

n-dodecane

2-methylpentadecane

3-methyldodecane

2,9-dimethyldecane

I d d l f l l h lImproved, new models for large alcohols:Improved models for large methyl esters:

Reduced models for fuel surrogates for CFD engine simulations:

Future challenges



Why do we need surrogate fuels?

• Real fuels are complex mixtures (100s of components)

n-paraffinsOxigenatesEtOH , MTBE, ETBE

oxygenates

• We cannot simulate all those components

• Define simpler mixtures that

arom aticsCH 3CH 3

represent properties of real fuels

=> surrogatesnaphthenes

Iso -paraffins

CH3CH3

olefins

Chemical classes in gasoline with grepresentative fuel components



Developed new procedure to formulate gasoline surrogates to match target gasoline fuels

Sl i NTC i l t t t iti it RON MON

Ignition delay at 825K correlates to octane number

1.0E‐01

25 atm T=825 Kslope

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

1 0E 03

1.0E‐02

on Delay Tim

es [s]

1.0E‐04

1.0E‐03

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5

Ignitio

1000K/T

40 different gasoline surrogates were simulated



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

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

16 140

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

10

12

14

16

PRFs

y

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

80

100

120

140b)

0 60 2 2 64 8 862

4

6

8

Higher unsaturated content

Sensitivity

40

60

80

PRFs

OI

y = ‐0.60x2 + 2.64x + 8.86R² = 0.79

‐2

0

‐3 ‐2 ‐1 0 1 2 3 4

Slope

0

20

‐3.5 ‐3 ‐2.5 ‐2 ‐1.5 ‐1Ignition Delay Times [Log(1/s)]

25 atm

Starting from the analysis of 40 different fuel mixtures a general approach to the formulation of the gasoline surrogate has been de eloped on the basis of the ke feat res of the ignition dela c r es



developed on the basis of the key features of the ignition delay curves

Matched gasoline properties with a surrogate

Surrogate (%Vol) RD387 Gasoline (%Vol)

Alkanes

i-Alkanes 0.57n- Alkanes 0.16Aromatics 0.23 0.227

Ol fi 0 04 0 042

0.731

Olefins 0.04 0.042A/F Ratio 14.61 14.79

H/C 1.925 1.946Octane index 87 87

Aromatics

Octane index 87 87Sensitivity 4 4

Olefins



Three different fuels were tested in the UCONN Rapid Compression Machine

RD387 Gasoline LLNL Surrogate Stanford A Surrogate Liquid Volume % of n-Alkanes

Liquid Volume % of Cyclo-Alkanes

Liquid Volume % of iso-Alkanes

aromatics aromaticsaromatics

cycloalkanes of iso-Alkanes

Liquid Volume % of Olefins

Liquid Volume % iso-alkanes iso-alkanes iso-alkanes

olefins olefins

of Aromatics

• The Stanford A Surrogate has been formulated targeting high temperature shock tube data

Gauthier, Davidson, Hanson, Combustion and Flame (2004)From: Kukkadapu, Kumar, Sung, Mehl, and Pitz, Combust. Flame (2012)



Comparison of experimentally-measured pressure histories in the RCM for 2 proposed surrogates and RD387 gasoline

30

35

40

PC=20 bar, TC=758 K

) Gasoline

Stanford A

30

35

40

PC=20 bar TC=663 Kr)

Stanford A

15

20

25

30

Pres

sure

(bar

)

LLNL

Gasoline

15

20

25

30 PC 20 bar, TC 663 K

Pres

sure

(bar

Gasoline

LLNL

10-10 -5 0 5 10

Time (ms)

10-10 0 10 20 30 40 50 60

Time (ms)

70

80Stanford A

70

80

Stanford A

40

50

60

70PC=40 bar, TC=693 K

Pres

sure

(bar

)

Gasoline

LLNL

40

50

60

70 PC=40 bar, TC=662 K

Pres

sure

(bar

)

LLNL

20

30

-10 -5 0 5 10 15Time (ms)

Gasoline

20

30

-10 0 10 20 30 40 50Time (ms)

Gasoline

Th LLNL t ll d th t t l i iti d l ti d th h t l



• The LLNL surrogate well reproduces the total ignition delay times and the heat release associated with the first stage ignition

From: Kukkadapu, Kumar, Sung, Mehl and Pitz, Proc. Combust. Inst.(2012)

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

100(a) =1.0, PC=20 bar

100(b) =1.0, PC=40 bar

10

lay

(ms)

10

100

y (m

s)

RCMRCM

1

Current Workal Ig

nitio

n D

el

1 Current WorkIgni

tion

Del

ay

Shock tube Shock tube

0.01

0.1

0 6 0 8 1 1 2 1 4 1 6

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

Tota

0.1

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

Tota

l 0.6 0.8 1 1.2 1.4 1.6

1000/TC (K-1)0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000/TC (K-1)

RC

Kukkadapu, Kumar, Sung, Mehl, and Pitz, Combust. Flame (2012)



CM

We are filling out components in the Diesel Fuel Surrogate Palette

n-hexadecane

n-alkanes

1-methylnaphthalene

n-dodecane

n alkanesbranched alkanescycloalkanesaromatics

tetralin

1,2,4-trimethylbenzene

2,9-dimethyldecane

2-methylpentadecane

decalin

n-dodecylcyclohexane

3-methyldodecane



Diesel surrogate components selected by AVFL-18

Have

Mueller, Cannella, Bruno, Bunting, Dettman, Franz, Huber, Natarajan, Pitz, Ratcliff and Wright, "Methodology for Formulating Diesel Surrogate Fuels with Accurate Compositional, Ignition-Quality, and Volatility Characteristics " Energy & Fuels (2012)



Characteristics, Energy & Fuels (2012)

Comprehensive kinetic model for aromatics

C0-C4 C0-C1

LLNL Galway/Princeton

C0-C4

Galway/Orleans/LLNL

C0-C4

Galway/LLNL

Version 49 (2009)

C0 C1

Version 54 (2012) Version 54 (2012)C2-C4

Cyclopentadiene

Benzene

Toluene

Cyclopentadiene

Benzene

Toluene

Cyclopentadiene

Benzene

Toluene

Cyclopentadiene

Benzene

TolueneToluene

Ethylbenzene

Xylenes

Toluene Toluene

n-Propylbenzene

n-Butylbenzene

Toluene

Ethylbenzene

Xylenes

n-Propylbenzenen-Butylbenzene n-Propylbenzene

n-Butylbenzene

Objective: a new, fully internally coherent and comprehensive mechanism for aromatic species to be used for future developments and surrogates’ simulations



p p g

The new mechanism core improves high temperature predictions (flame speeds, ignition, …)

Chemical kinetic model validation• Idealized chemical reactors with/without simplified transport

phenomenonCombustion ParametersSh k t bJet Stirred Reactors Premixed Laminar FlamesCombustion Parameters

Temperature

Pressure

Mixture fraction (air-fuel ratio)

Shock tube

Non Premixed Flames

( )

Mixing of fuel and air

Rapid Compression Machine

Engine

Rapid Compression Machine

Engine CombustionElectric Resistance

Heater

Evaporator

Fuel Inlet

Wall Heaters

High pressure flow reactors



Slide Table

Oxidizer Injector

Optical Access Ports

Sample Probe

Validation of comprehensive aromatic model

80

100CyclopentadieneNaphtaleneBenzeneIndenect

ivity

1000

10000PHI 0.5PHI 1.0PHI 2.0es

[µs]

Orme, J.; Simmie, J. M. 2005Kim, D. H.; Mulholland, J. A.; Wang, D.; Violi, A 2010

Shock tube

Flow reactor

20

40

60

IndeneStyrene

% C

arbo

n Se

lec

PAH

growth 100

1000

nitio

n D

elay

TimFlow reactor

Pyrolysis

P= 1 atm

0

20

800 900 1000 1100 1200 1300Temperature [K]

105 5.4 5.8 6.2 6.6 7 7.4 7.8

10000K/T[K]

Ig

60

1% C5H6 in Ar, 3.8 atm

Ji, Egolfopoulos, et al. PCI 2012, CnF20127 0E 041 2E 04

Alzueta, M. U.; Glarborg, P.; Dam-Johansen, 2000

40

50

60

d [c

m/s

]

g p

4.0E‐04

5.0E‐04

6.0E‐04

7.0E‐04

8.0E‐05

1.0E‐04

1.2E‐04

Frac

tion

CO

,CO

2 MCO2

C6H6

Flow reactor

10

20

30Fl

ame

Spee

d

1.0E‐04

2.0E‐04

3.0E‐04

2.0E‐05

4.0E‐05

6.0E‐05

C6H

6M

ole

F Mole Fraction

CO

Φ=0.62

P= 1.05 atmFlame speed

T=353K



00.6 0.8 1 1.2 1.4 1.6

Φ

0.0E+000.0E+00900 1000 1100 1200 1300 1400 1500

Temperature [K]

T 353K

Progress on kinetic models for lightly-methylated iso-alkanes (present in diesel, jet, and hydrotreated renewable fuels)

50

55 air, 353 K, 1atm experiments (symbols), model (lines)

t2-methylheptane

40

45

50

Suo (cm/s)

n-octane

3-methylheptaneed [c

m/s

]

30

35

r Flame Speed,

S

n‐octane

2‐methylheptane 2,5-dimethylhexane Experiments performed at

USC, Los Angeles

3-methylheptane

r fla

me

spee

15

20

25

Laminar 3‐methylheptane

2,5‐dimethylhexane

Experiments: symbols

Model: lines

Lam

inar

(LLNL)

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

Important to properly predict flame speeds for gas turbine applications

15 0.6 0.8 1.0 1.2 1.4 1.6

Equivalence Ra o Equivalence ratio



Location of single methyl branch has minimal effect on flame speed

Biofuels

rapeseed

Camelina SativaCamelina Sativa

Sustainable Feedstock algal pilot scale bioreactor

sugarcane



sugarcane

For the first time, experimental validation data are available for main components in biodieselp

Methyl Palmitate (C16:0)

triglycerideOO

O

O

R R

+ 3 CH 3OH

Fatty acid methyl esters (FAMEs):

Methyl Stearate (C18:0)

methanolO

O

O

R OHO

70

Methyl Oleate (C18:1)methyl ester glycerol

OHOH

CH3OR

3 +

3040506070

%

SoybeanRapeseed Methyl Linoleate (C18:2)

0102030

Methyl Linolenate (C18:3)



C16:0 C18:0 C18:1 C18:2 C18:3

Methyl oleate model compares well to shock tube ignition experiments at elevated pressures

methyl oleate:yReal biodiesel methyl ester

3.5 atm

7 atm

Lines: LLNL model

Squares: Stanford measurementsStanford aerosol shock tubeStanford aerosol shock tube



Campbell, Davidson, Hanson and Westbrook, Proc. Combust. Inst (2012)

Chemical kinetic models for other biofuels

isopentanol• Exhibits intermediate temperature heat release needed to• Exhibits intermediate temperature heat release needed to

obtain high load operation with HCCI (Dec and Yang, 2010) 1-pentanol

Higher energy density than ethanol• Higher energy density than ethanol 1-butanol

• Has higher energy density than ethanol and can be mixed in fuel pipelines. Can be made from renewable sources

diethylcarbonate• Sugar cane in Colombia is converted to ethanol. The ethanol

and CO2 products are converted to diethyl carbonate. Diesel replacement fuel. (Hosted visiting student from Columbia.)



Iso-pentanol ignition in a shock tube and rapid compression machine

1000Isopentanol, Phi=1.0, P=0.7‐2.3 MPa, in air

100

ay [m

s]

Curves: LLNL modelSymbols: NUIG experimental data R

CM

1

10

nitio

n de

la

ST Exp.: P=0.7 MPaCal.: P=0.7 MPaRCM Exp : P=0 7 MPa

Rapid compression machineExperimental data: UConn

RCM7 bar M

0.1

Ign RCM Exp.: P=0.7 MPa

Cal. w heat loss.: P=0.7 MPaST Exp.: P=2 MPaCal.: P=2 MPaRCM Exp.: P=2 MPaCal w heat loss : P=2 MPaShock tube Shock tube

20 bar

0.017 9 11 13 15

10000/T [1/K]

Cal. w heat loss.: P 2 MPaExperimental data: NUIG



From: Tsujimura, Pitz, Gillespie, Curran, Weber, Zhang and Sung, Energy & Fuels (2012)

Shock tube ignition delay times for 1-butanol at engine-like pressures and temperatures

=1 in air 1 in air

20 atm 80 atmShock tube

10 atm 30 atm

Curves: LLNL model

S b l E i t f RWTH A h U i itSymbols: Experiments from RWTH Aachen University

F S th V k Y M hl Oß ld M t lf W tb k Pit K h Höi h F d


Lawrence Livermore National LaboratoryExperiments from Vranckx et al., Comb. Flame 2011 and Heufer et al. Proc. Combust. Inst. 2010

From: Sarathy, Vranckx, Yasunaga, Mehl, Oßwald, Metcalfe, Westbrook, Pitz, Kohse-Höinghaus, Fernandes, and Curran, Combust. Flame (2012)

Mechanisms are available on LLNL website and by email

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


Lawrence Livermore National LaboratoryLLNL-PRES -427942

Summary: Improved/new models for many different fuel components

Improved/new models for • AromaticsAromatics• Methyl esters• Alcohols

Butanol Iso-pentanol n-pentanol

Reduced models for use in CFD engine codes:

n-dodecane



Challenges for kinetic model development for surrogate fuels for IC engines: Higher pressures

Need experimental validation data at 100-200 bar with EGR (dilution by H2O and CO2) for advanced diesel engine combustion regimes

Reactivity Controlled Compression Ignition (RCCI) Engine:

Ignition near 130 bar



Reitz et al., 2012, www.annualmeritreview.energy.gov

Documents

Lawrence Livermore National Laboratory · USC, Los Angeles r flame spe 15 20 Lamina 25 3‐methylheptane 2,5‐dimethylhexane Experiments: symbols Lamina Model: lines (LLNL) (USC)