Detailed Kinetic Modeling of Gasoline Fuels

1

LLNL-PRES-653562

Detailed Kinetic Modeling

of Gasoline Fuels

Slide show prepared by Dr. Marco Mehl, LLNL

Charles Westbrook

Tsinghua-Princeton 2014 Summer School on

CombustionJuly 20-25, 2014

2

LLNL-PRES-653562

What is the relationship between fuel quality and

engine technology?

3

LLNL-PRES-653562

The SI Engine

0

0.2

0.4

0.6

0.8

1

0

5

10

15

20

25

335 360 385 410

Pres

sure

[bar

]

Crank Angle Degree

Bur

nt m

ass

frac

tion

The spark starts the propagation of a turbulent flame frontControlled Temperature and Pressure increase

Nikolaus August Otto (1832–1891)

Motored Cycle

Combustion cycle

Otto Engine (1876)

4

LLNL-PRES-653562

The first car road-trip in history (5th August 1888)

Bertha Benz (1849–1944)

The Driver Her Car

Patent-Motorwagen No. 3

The Route

Mannheim to Pforzheim and back (194 km)

The Fuel

Petroleum Ether

The First Filling Station

Willi Ockel’s Pharmacy, Wiesloch

5

LLNL-PRES-653562

Early fuels for transportation 1885-1910

Petroleum Ether

Pioneers of the car industry used the fuels available at the time.

The best fit to their need was a light fraction of crude oil distillation: a byproduct of the production of lamp oil

This distillate was mainly composed of n-alkanes (C5-C7) and was sold to be used as a solvent

Its availability was limited (10-20% yield from crude oil fractional distillation)

6

LLNL-PRES-653562

0

50

100

150

200

250

300

350

0

2

4

6

8

10

12

14

16

1880 1900 1920 1940 1960 1980 2000 2020

Increasing power output required bigger engine and higher compression ratio…

20%

30%

40%

50%

60%

70%

0 2 4 6 8 10 12 14 16

Theoretical Thermal efficiency

Compression Ratio

Year

Power (H

p)C

ompr

essi

on R

atio

7

LLNL-PRES-653562

But soon engine designers (and pilots) faced catastrophic failures….

Engine became noisy and, in some cases, the cylinders it

literally blew up!

8

LLNL-PRES-653562

When the compression ratio was increased anomalous combustion behavior lead to engine failure

0

0.2

0.4

0.6

0.8

1

0

5

10

15

20

25

335 360 385 410

Pres

sure

[bar

]

Crank Angle Degree

Bur

nt m

ass

frac

tion

Knocking Cycle

9

LLNL-PRES-653562

H

•

R•

- RH

What causes knock?

Rea

ctiv

ityReactor Temperature

•

Fast High

Temperature Combustion

Hi T

Mechanism

Straight Chain Alkanes

NTC

+ O2

Degenerate Branching Path

OO•

OOH•

•OO

OOH

HOO

O

+ O2

- •OHO

O

•

•OH+ +

Low T

Mechanism

+ HO2•

+ •OH

+ •OHO

+

O

HO2 Radicals lead to the formation of H2O2, when the temperature increases H2O2 2OH triggering the transition to high temperature ignition

10

LLNL-PRES-653562

Standard tests to measure the autoignition propensity of the fuel were introduced

RON: - Speed 600 rpm

- Inlet Temperature 325K

- Spark Advance 13°

Representative of idling/low load conditions

MON: - Speed 900 rpm

- Inlet Temperature 422K

- Spark Advance 19-26°

Representative of road/high load conditions

The critical compression ratio (when knock reaches a prefixed intensity in a variable CR engine) was used to compare the fuel candidate to a scale of reference fuels constituted by n-heptane (PRF0) and iso-octane (PRF100)

CFR Engine (1929)

It should be noted that today’s octane rating are still based on the same type of tests

11

LLNL-PRES-653562

Thomas Midgley demonstrated tetraethyl lead to be an effective anti-knock agent (despite its toxicity)

H

•

R•

- RH

•

Fast High


+ O2


OO•

OOH•

•OO

OOH

HOO

O

+ O2

- •OHO

O

•

•OH+ +

+ HO2•

+ •OH

+ •OHO

+

O

It is believed that TEL forms a “fog” of PbO which scavenges HO2 radicals

and prevents the build up of H2O2inhibiting the transition to high

temperature ignition

0.0001

0.001

0.01

0.1

0.8 1 1.2 1.4 1.6

Ignitio

n de

lay tim

e [s]

1000K/T [K]

+ PbO

25 atm, Φ=1, constant volume

Pitz & Westbrook, Combustion and Flame 63: 113-133 (1986)

12

LLNL-PRES-653562

The high demand for gasoline drove the quest for new refinery technologies

1920 circa: Thermal Cracking:

• Decomposition of heavier petroleum fractions to shorter (more volatile) alkanes and olefins

• Twofold increase of the yield of gasoline from oil

1937: Catalytic Cracking:

Higher yields, better fuels (higher aromatic and lightly branched paraffins)

13

LLNL-PRES-653562

The high demand for gasoline drove the quest for new refinery technologies

1949: Reforming:

Upgrade the fuel quality by increasing the fraction of aromatics and branched paraffins

14

LLNL-PRES-653562

0.0001

0.001

0.01

0.1

0.8 1 1.2 1.4 1.6

Ignitio

n de

lay tim

e [s]

1000K/T [K]

Alkenes are less reactive at LT than their saturated homologues due to alternative propagation pathways and weaker R-OO bond

•

R• H

- RH

+ O2


OO•

OOH•

•OO

OOH

HOO

O

+ O2

- •OHO O

••OH+ +

+ HO2•

+ •OH

+ •OHO

+

O

•

HTHR

H

OH

•

+ O2

H

OH

OO•

OH

•

+ •OH+O

O

Non- Branching Low Temperature Pathways


15

LLNL-PRES-653562

0.0001

0.001

0.01

0.1

0.8 1 1.2 1.4 1.6

Ignitio

n de

lay tim

e [s]

1000K/T [K]

If the abstraction reaction involves a tertiary site of an isoalkanethe formation of ketohydroperoxides is not possible

H

•

R•

- RH

•

Fast High


+ O2OO•

OOH•

•OO

OOH

+ O2

- •OH

+ HO2•

+ •OH

+ •OHO

+

O

Ketohydroperoxide

The internal isomerization steps required by the low temperature

degenerate branching mechanism are slower because of the higher

activation energy required for primary hydrogens


16

LLNL-PRES-653562

0.0001

0.001

0.01

0.1

0.8 1 1.2 1.4 1.6

Ignitio

n de

lay tim

e [s]

1000K/T [K]

The 3D structure of cycloalkane makes H atom unavailable for the internal abstraction reactions reducing the reactivity of the fuel at low temperature


The presence of side chains, on the other hand, can provide

alternative abstraction sites that increase significantly the

reactivity of the cycloalkane molecule

17

LLNL-PRES-653562

0.0001

0.001

0.01

0.1

0.8 1 1.2 1.4 1.6

Ignitio

n de

lay tim

e [s]

1000K/T [K]

Aromatics don’t have LT degenerate branching path and act as radical scavengers in mixtures (stable resonantly stabilized benzyl radicals)

Hi T (O and H radical formation)

Intermediate and low T (HO2radicals, resonantly stabilized radicals and termination reactions


18

LLNL-PRES-653562

The introduction of additives, new refining technologies and fuel standards improved fuel quality, thus enabling the development of modern engines

0

50

100

150

200

250

300

350

0

2

4

6

8

10

12

14

16

1880 1900 1920 1940 1960 1980 2000 2020

Direct injection

1924 TEL (GM)

1928 Thermal Cracking 1929 Octane Rating Tests

TELOxygenates

1937 Catalytic Cracking

1949 Reforming

Environmental standards

Year

Power (H

p)C

ompr

essi

on R

atio

19

LLNL-PRES-653562

How does gasoline look like today?

Today gasolines are complex mixtures of hundreds of components

Qualitative composition is known

Their composition is intrinsically variable (seasonally, depending on the feedstock,

….)

The introduction of alternative fuels is making things even more complicated

A detailed kinetic combustion model can include hundreds of species to describe a

single component

Simpler mixtures have to be defined for modeling and more reproducible

experiments

Gasoline 6 < C < 10 Jet fuel 9 < C < 13 Diesel 13 < C < 22

20

LLNL-PRES-653562

Surrogate fuels allow to cope with the compositional complexity of real fuels

• Mixtures with a limited number of components (Surrogates) are generallyused to model the physical and chemical properties of real fuels both inmodeling and experimental studies (simplicity & consistency)

• The simplest surrogate for gasoline is considered to be iso-octane

• Primary reference fuels allow to characterize better the auto-ignition propensity allowing a finer tuning of the reactivity. The notion of AKI = (RON+MON)/2 (currently used in the US as a “road octane number”) was introduced and PRFs with that rating were used to simulate gasoline

• The shortcomings of these surrogates become evident when we are moving away from the operating conditions of old carbureted engines (DISI, Downsized, HCCI, …)

21

LLNL-PRES-653562

A numerical study on the engine combustion behavior of an olefinicgasoline

Type Spark Ignition Valves/cylinder 4

Stroke 78.9 mm Bore 70.8 mm

Compresion Ratio 10.6 Maximum Power 51.5 kW @ 5000 rpm Maximum Torque 104.5 Nm @ 4000 rpm

FIAT-Lancia 1200 16v

An “Old School” engine

Engine schematization in GASDYN (Politecnico di Milano) a 1-D/quasi-D model

22

LLNL-PRES-653562

Engine Thermo Fluid Dynamic Modeling

All the typical boundaries encountered in i.c. engine duct systems (valves, abruptarea changes, throttle valves, turbines, compressors …) are modeled adopting theclassical assumption of quasi-steady flow.

A. Onorati, G. Ferrari, G. Montenegro, A. Caraceni, P. Pallotti, Prediction of S.I. Engine Emissions during an ECE Driving Cycle via Integrated Thermo-Fluid Dynamic Simulation, SAE Int. Congress & Exp. (Detroit) 2004, paper n. 2004-01-1001

www.engines.polimi.it

Burned ZoneEquilibrium Conditions

Unburned ZoneDetailed Chemistry

burned

unburned

burned

unburned

Burning Rate estimated with Fractal Approach

(minimum flame wrinkling is proportional to the Kolmogorov scale, the maximum one to the integral length)

tt l

l

Au =u

A

23

LLNL-PRES-653562

Engine Model Validation100

50

60

70

80

90

1000 2000 3000 4000 5000 60001000 2000 3000 4000 5000 6000

Experimental

CalculatedVolu

met

ric E

ffic

ienc

y [%

]

Engine Speed [rpm]

100

50

60

70

80

90

1000 2000 3000 4000 5000 60001000 2000 3000 4000 5000 6000

Experimental

CalculatedVolu

met

ric E

ffic

ienc

y [%

]

Engine Speed [rpm]

50.0

70.0

90.0

110.0

130.0

1000 2000 3000 4000 5000 60001000 2000 3000 4000 5000 6000

Engine Speed [rpm]

Indi

cate

d To

rque

[Nm

]

Experimental

Calculated

50.0

70.0

90.0

110.0

130.0

1000 2000 3000 4000 5000 60001000 2000 3000 4000 5000 6000

Engine Speed [rpm]

Indi

cate

d To

rque

[Nm

]

Experimental

Calculated

Calculated vs Experimental Volumetric Efficiency, Torque and Pressure cycle

0

10

20

30

40

50

60

70

0 50 100 150 200 250 300 350Volume [cm3]

Pres

sure

[bar

]

Calculated, SA 8° BTDC


Experimental, SA 8° BTDC

80

Calculated Knocking CycleSA 26° BTDC

0

10

20

30

40

50

60

70

0 50 100 150 200 250 300 350Volume [cm3]

Pres

sure

[bar

]



Experimental, SA 8° BTDC

80

Calculated Knocking CycleSA 26° BTDC

24

LLNL-PRES-653562

Gasoline SurrogatesSurrogate 1 Surrogate 2

Density (15°C) Kg/m3 0.757 0.750Reid Vapour Pressure kPa 21.0 19.5RON - 97.3 97.3MON - 89.2 86.6

IBP °C 66 7610% °C 89 8950% °C 99 9990% °C 102 102FBP °C 108 109

n-heptane vol % 13 19iso-octane vol % 42 24Toluene vol % 32 26MTBE vol % 13 13di-isobutylene vol % - 18

25

LLNL-PRES-653562

-18 -16 -14 -12 -10

1.43

0.99

1.10

1.24

14-1612-1410-128-106-84-62-40-2

-18 -16 -14 -12 -10

1.43

0.99

1.10

1.24

14-1612-1410-128-106-84-62-40-2

Spark Advance

Φ

Unburnt Fraction at Knock vs Engine Operating Conditions

(4000 rpm, full load)

Knock Critical Conditions assuming 10% of unburned gases as “light knock” threshold

We define the octane performance based on the PRF showing the same knock propensity

0.5

1

1.5

-18 -16 -14 -12 -10 -8

Knock Region

Spark Advance BTDC

0.5

1

1.5

-18 -16 -14 -12 -10 -8

Knock Region

Spark Advance BTDC

ExperimentalSurrogate Calculation88 PRF Calculation

?

0.5

1

1.5

-18 -16 -14 -12 -10 -8

Knock Region

Spark Advance BTDC

0.5

1

1.5

-18 -16 -14 -12 -10 -8

Knock Region

Spark Advance BTDC

ExperimentalCalculation88 PRF Calculation

Φ

Knock Limits

26

LLNL-PRES-653562

FIAT-Lancia ComparisonsExperimental and CalculatedOctane Performances ofdifferent fuels in on-roadconditions…

Effect of alkenes:Engine Octane RequirementGasoline without AlkenesGasoline containing Alkenes

80

83

86

89

92

3000 3500 4000 4500 5000

Surrogate 1Surrogate 2Engine

Engine speed (rpm)

Ben

ch O

N a

nd o

ctan

e re

quire

men

t (P

RF’

s)

27

LLNL-PRES-653562

Effect of Turbocharging

Turbocharged conditions highlightthe better octane performances ofsurrogate 2 (containing alkenes)

TurboCharged Engine

The lower reactivity of alkenes at lowT and high P allows higher pressureboost.

2000 2500 3000 3500 4000 4500 5000

Benc

hO

N (P

RF’s

)

86

88

90

92

94

96

Engine speed [rpm]

b)

Surrogate 1

Surrogate 2

2000 2500 3000 3500 4000 4500 5000

Benc

hO

N (P

RF’s

)

86

88

90

92

94

96

Engine speed [rpm]

b)

2000 2500 3000 3500 4000 4500 5000

Benc

hO

N (P

RF’s

)

86

88

90

92

94

96

Engine speed [rpm]2000 2500 3000 3500 4000 4500 5000

Benc

hO

N (P

RF’s

)

86

88

90

92

94

96

Engine speed [rpm]

b)

Surrogate 1

Surrogate 2Surrogate 1Surrogate 2

1.0

1.2

1.4

1.6

1.8

criti

cal a

bsol

ute

pres

sure

(bar

)

2000 2500 3000 3500 4000 4500 5000

Engine speed [rpm]

a)

Surrogate 1

Surrogate 2

1.0

1.2

1.4

1.6

1.8

criti

cal a

bsol

ute

pres

sure

(bar

)

2000 2500 3000 3500 4000 4500 5000

Engine speed [rpm]

a)1.0

1.2

1.4

1.6

1.8

criti

cal a

bsol

ute

pres

sure

(bar

)

2000 2500 3000 3500 4000 4500 5000

Engine speed [rpm]

1.0

1.2

1.4

1.6

1.8

criti

cal a

bsol

ute

pres

sure

(bar

)

2000 2500 3000 3500 4000 4500 5000

Engine speed [rpm]

a)

Surrogate 1

Surrogate 2

Surrogate 1Surrogate 2

28

LLNL-PRES-653562

600 650 700 750 800 850 900 950 1000

10

20

30

40

50

60

70

T [K]

Pre

ssur

e[b

ar]

600 650 700 750 800 850 900 950 1000

10

20

30

40

50

60

70

T [K]

Pre

ssur

e[b

ar]

MON

RON

Super-charged

Carbureted

Fuel Injected

Modified from Andy D. B. Yates, André Swarts and Carl L. Viljoen, Correlating Auto-Ignition Delays And Knock-Limited Spark-Advance Data For Different Types Of Fuel, SAE 2005-01-2083

New engine technologies shift the operating region of the engine toward lower temperature and higher pressures

HCCI (NA to Boosted)

29

LLNL-PRES-653562

T [K]

600 650 700 750 800 850 900 950 10001

10

20

30

40

50

P [B

ar]

T [K]

600 650 700 750 800 850 900 950 10001

10

20

30

40

50

P [B

ar]

Induction times in constant volume systems and reactivity maps

Reactivity of the system is expected to increase with pressure andtemperature

30

LLNL-PRES-653562

600 650 700 750 800 850 900 950 1000

10

20

30

40

50

60

70

Pre

ssur

e [b

ar]

T[K]

600 650 700 750 800 850 900 950 1000

10

20

30

40

50

60

70

Pre

ssur

e [b

ar]

T[K]

Surrogates Reactivity

Critical Operating Conditions in Carbureted and Turbo Engines

The different shape of the NTC region explains the different performances of the two surrogates

Surrogate 1

Surrogate 2

31

LLNL-PRES-653562

More complex surrogates are needed for today’s engines – Can we use a kinetic model to assist the development of better surrogates???

• As we have seen, engine technology and fuel quality has changed substantially from the ‘20s

Are the standards defined back then, based on the ignition behavior of PRFs, still relevant to the combustion behavior of gasoline in modern engines?

• Mixtures of saturated, unsaturated, oxygenated, aromatic compounds are needed to reproduce correctly the reactivity of the fuel

• A predictive model should take into account fuel effects on combustion and emissions

32

LLNL-PRES-653562

We can use a broadly validated kinetic mechanism to analyze the combustion behavior of gasoline surrogates

0.1

1

10

100

50 atm

41 atm

20 atm

6.5 atm10 atm

13 atm

3-4.5 atm

30 atm

n-heptane

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5

1000K/T

Igni

tion

Del

ay T

imes

[ms]

1

Igni

tion

Del

ay T

imes

[ms]

toluene

0.7 0.8 0.9 1 1.1

1000K/T

Shock Tube 12 atm

Shock Tube 55 atm

Rapid Compression Machine 44 atm

10

1000

100

0.1

0.010.1

1

10

100

1-hexene

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000K/T

8.6-10.9 atm

6.8-8.4 atm7.7-9.6 atm

Igni

tion

Del

ay T

imes

[ms]

1

10

100

1000

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000K/T

Igni

tion

Del

ay T

imes

[ms]

Toluene 15 atm

Mixture 3.9-4.9 atm

N-heptane 3.3-4.6 atm

1

10

100

1000

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000K/T

Igni

tion

Del

ay T

imes

[ms]

Toluene 15 atm

Iso-octane 12.6-16.1 atm

Mixture 12-14.6 atm

1

10

100

1000

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000K/T

Igni

tion

Del

ay T

imes

[ms]

Toluene 15 atm

1-hexene 8.5-10.9 atm

Mixture 11.4-13.9 atm

1

10

100

1000

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000K/T

Igni

tion

Del

ay T

imes

[ms]

1-hexene 8.5-10.9 atm

Iso-octane 12.6-16.1 atm

Mixture 11.4-14 atm

1

10

100

1000

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

1000K/T

Igni

tion

Dela

y Tim

es [m

s]

Surrogate 11.8-14.8 atm

47% 35% 18%

50% 50%

65% 35%

30% 70%

82% 18%

33

LLNL-PRES-653562

How does the presence of aromatics, olefins and oxygenates influence the reactivity of a fuel?

1.00E-04

1.00E-03

1.00E-02

1.00E-01

0.8 1 1.2 1.4 1.6

1000K/T

Igni

tion

Del

ay T

imes

[s]

1.00E-04

1.00E-03

1.00E-02

1.00E-01

0.8 1 1.2 1.4 1.6

AKI=85, SEN=15

The two octane numbers (RON and MON) are used to determine theAKI=(RON+MON)/2 and the sensitivity (RON-MON) of the fuel to theoperating conditions

AKI=0, SEN=0

AKI=85, SEN=3

AKI=100, SEN=0

34

LLNL-PRES-653562

40 different surrogates were simulated using the detailed kinetic mechanism and their octane performances analyzed

Sensitivity vs. Slope AKI vs. Ignition Delay time @ 825K

Two correlations between the octane numbers and the autoignition behavior in a kinetically controlled system were developed

Data: Personal communication, N.Morgan et al. Comb. & Flame

‐2

0

2

4

6

8

10

12

14

16

‐3 ‐2 ‐1 0 1 2 3 4

PRFs

Higher unsaturated content

Slope NTC

Senitiv

ity

a)

0

20

40

60

80

100

120

140

‐3.5 ‐3 ‐2.5 ‐2 ‐1.5 ‐1

PRFs

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

I

b)

35

LLNL-PRES-653562

Boosted HCCI for High Power without Engine Knock and with Ultra-Low NOx Emissions – using Conventional Gasoline (Dec and Yang)

Approach:Conventional Gasoline: (R+M)/2 = 87, RON = 90.8, MON = 83.2. Aromatics

23%, Olefins 4.2%, Alkanes 73%Piston: CR =14, open combustion chamber.Current data at 1200 rpm.

Control pressure-induced enhancement of autoignition with a combination of:

• Intake temperature control• Cooled EGR

Late combustion phasing thanks to intermediate temperature heat release

36

LLNL-PRES-653562

To understand the chemistry underlying this behavior we need a surrogate for the RD387 full blend research gasoline.

Gasoline

Alkanes 73.1 %Vol

Aromatics 22.7 %Vol

Olefins 4.2 %Vol

H/C 1.946022

RON 83.2

MON 90.8

4 Component surrogate:

iso-octane

n-heptane

toluene

2-pentene

Base components of PRFs

High Sensitivity, best validation

High Sensitivity, molecular weight in the gasoline range

Limited information on the composition of the gasoline is typically available

We have a well validated model for these 4 components

We start determining the fuel component palette

37

LLNL-PRES-653562

A numerical approach to the formulation of gasoline surrogates

The aromatic and olefin content is matched to the target (H/C ratio)

and a preliminary surrogate is proposed

The ignition delay time curve is calculated

The composition is adjusted

The correlations are used to estimate the AKI and SEN of the

surrogate model

‐2

0

2

4

6

8

10

12

14

16

‐3 ‐2 ‐1 0 1 2 3 4Slope NTC

Senitiv

ity

a)

0

20

40

60

80

100

120

140

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

AKI

b)

Validation (if available)

Engine Calculations

NO

YES

Does the surrogate match the target?

The target for our fuel surrogate formulation is the overall composition, the H/C ratio and the autoignitionbehavior

38

LLNL-PRES-653562

The calculated ignition delay curve of the surrogate is compared with ignition delay data of RD387

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

Ove

rall

Igni

tion

Del

ay (m

s)

1000/TC (K-1)

=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


Ove

rall

Igni

tion

Del

ay (m

s)

1000/TC (K-1)

=0.5, PC=20 bar

0.01

0.1

1

10

100

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6


Ove

rall

Igni

tion

Del

ay (m

s)

1000/TC (K-1)

=0.5, PC = 40 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

Ove

rall

Igni

tion

Del

ay (m

s)

1000/TC (K-1)

=1.0, PC=40 bar

The fuel surrogate formulated based on the theoretical assumptions showed a very good agreement with gasoline data acquired in Rapid Compression Machine and Shock Tube experiments

39

LLNL-PRES-653562

HCCI simulations have been performed to investigate the chemistry responsible for the ITHR (SAE 2010-01-1086)

369

370

371

372

373

374

375

376

377

50 100 150 200 250 300 350

Calc CA50

Exp CA10

Pin kPa

CA

D

Pin kPa

350

360

370

380

390

400

410

420

430

50 100 150 200 250 300 350 400

T @ BDC (From exp T profile) [K]

T @ BDC (Cv corrected) [K]

Surrogate T @ BDC [K]

T @

BD

C

EGR

The intake temperature necessary to achieve the desired combustion timing is successfully reproduced

Heat release profiles are well reproduced too

-35 -25 -15 -5 5

HRR 100kPa Norm

HRR 130kPa Norm

HRR 160kPa Norm

HRR 180kPa Norm

HRR 200kPa Norm

HRR 240kPa Norm

HRR 324kPa Norm

Gasoline: Experiments Surrogate: Calculations

0

0.002

0.004

0.006

0.008

0.01

-35 -25 -15 -5 5

Hea

t-R

elea

se R

ate

/ To

tal H

R [1

/°C

A]

Crank Angle relative to Peak HRR [°CA]

Pin = 325 kPa

Pin = 240 kPa

Pin = 200 kPa

Pin = 180 kPa

Pin = 160 kPa

Pin = 130 kPa

Pin = 100 kPa

b.

-35 -25 -15 -5 5Crank Angle relative to Peak HRR [°CA]

40

LLNL-PRES-653562

Kinetic Analysis

Crank Angle relative to Peak HRR [ºCA]

Nor

mal

ized

HR

R

0

0.0002

0.0004

0.0006

0.0008

0.001

-35 -25 -15 -5 5

HRR 100kPa Norm

HRR 130kPa Norm

HRR 160kPa Norm

HRR 180kPa Norm

HRR 200kPa Norm

HRR 240kPa Norm

HRR 324kPa Norm

OH radical flux analysis

(13 CAD BTDC 328kPa)

What is triggering the ITHR?

OH radical flux analysis

HRR reaction analysis

What is contributing to the ITHR?

Reaction contribution to the HRR

41

LLNL-PRES-653562

OH radical fluxes @ 13CAD BTDC

Low temperature branching from linear

alkanes

Non branching low temperature oxidation of the olefins (net flux nearly = 0)

42

LLNL-PRES-653562

Reaction Contribution to HRR

H+O2(+M)<=>HO2(+M)

2HO2 H2O2+O22OH

Methyl Oxidation to Formaldehyde via CH3O2H

Formaldehyde Oxidation (CH2O abstraction and HCO stabilization to CO+HO2)

Nor

mal

ized

HR

R

0

0.0002

0.0004

0.0006

0.0008

0.001

-35 -25 -15 -5 5

HRR 100kPa Norm

HRR 200kPa 418K Norm HRR 200kPa 377K Norm HRR 180kPa 345K Norm

Crank Angle relative to Peak HRR [ºCA] CAD ATDC

HR

R [E

rg/C

AD

]

‐1.0E+09

‐5.0E+08

0.0E+00

5.0E+08

1.0E+09

1.5E+09

2.0E+09

2.5E+09

3.0E+09

‐15 ‐10 ‐5 0 5 10

HRR 100kPa

43

LLNL-PRES-653562


H+O2(+M)<=>HO2(+M)

2HO2 H2O2+O22OH



Nor

mal

ized

HR

R

0

0.0002

0.0004

0.0006

0.0008

0.001

-35 -25 -15 -5 5

HRR 100kPa Norm



HR

R [E

rg/C

AD

]

‐1.0E+09

‐5.0E+08

0.0E+00

5.0E+08

1.0E+09

1.5E+09

2.0E+09

2.5E+09

3.0E+09

‐15 ‐10 ‐5 0 5 10

HRR 100kPa

‐1.00E+09

‐5.00E+08

0.00E+00

5.00E+08

1.00E+09

1.50E+09

2.00E+09

2.50E+09

3.00E+09

‐15 ‐10 ‐5 0 5 10

HRR 200kPa 418K

44

LLNL-PRES-653562


H+O2(+M)<=>HO2(+M)

2HO2 H2O2+O22OH



Nor

mal

ized

HR

R

0

0.0002

0.0004

0.0006

0.0008

0.001

-35 -25 -15 -5 5

HRR 100kPa Norm



HR

R [E

rg/C

AD

]

‐1.0E+09

‐5.0E+08

0.0E+00

5.0E+08

1.0E+09

1.5E+09

2.0E+09

2.5E+09

3.0E+09

‐15 ‐10 ‐5 0 5 10

HRR 100kPa

‐1.00E+09

‐5.00E+08

0.00E+00

5.00E+08

1.00E+09

1.50E+09

2.00E+09

2.50E+09

3.00E+09

‐15 ‐10 ‐5 0 5 10

HRR 200kPa 418K

‐1.00E+09

‐5.00E+08

0.00E+00

5.00E+08

1.00E+09

1.50E+09

2.00E+09

2.50E+09

3.00E+09

‐15 ‐10 ‐5 0 5 10

HRR 200kPa 377K

45

LLNL-PRES-653562


H+O2(+M)<=>HO2(+M)

2HO2 H2O2+O22OH



Nor

mal

ized

HR

R

0

0.0002

0.0004

0.0006

0.0008

0.001

-35 -25 -15 -5 5

HRR 100kPa Norm



HR

R [E

rg/C

AD

]

‐1.0E+09

‐5.0E+08

0.0E+00

5.0E+08

1.0E+09

1.5E+09

2.0E+09

2.5E+09

3.0E+09

‐15 ‐10 ‐5 0 5 10

HRR 100kPa

‐1.00E+09

‐5.00E+08

0.00E+00

5.00E+08

1.00E+09

1.50E+09

2.00E+09

2.50E+09

3.00E+09

‐15 ‐10 ‐5 0 5 10

HRR 200kPa 418K

‐1.00E+09

‐5.00E+08

0.00E+00

5.00E+08

1.00E+09

1.50E+09

2.00E+09

2.50E+09

3.00E+09

‐15 ‐10 ‐5 0 5 10

HRR 200kPa 377K

‐1.0E+09

‐5.0E+08

0.0E+00

5.0E+08

1.0E+09

1.5E+09

2.0E+09

2.5E+09

3.0E+09

‐15 ‐10 ‐5 0 5 10

H+O2(+M)<=>HO2(+M)

2HO2‐‐>H2O2+O2‐‐>2OH



R+O2 (R = fuel radical or primary fuel fragment, e.g. TC4H9 )

Low T HR

HRR 180kPa 345K: Hi Boost, Low Tin

HR

R [E

rg/C

AD

]

Crank Angle relative to Peak HRR [ºCA]

46

LLNL-PRES-653562

350

360

370

380

390

400

410

420

430

50 100 150 200 250 300 350 400

T @ BDC (From exp T profile) [K]T @ BDC (Cv corrected) [K]Surrogate T @ BDC [K]PRF 87 T @ BDC [K]

PRF87 Calculations vs. Surrogate Calculations

Using a PRF87 a much lower T @ BDC is required to match the same combustion phasing

A correct surrogate formulation is mandatory to model these data!

369

370

371

372

373

374

375

376

377

50 100 150 200 250 300 350

Exp CA10Calc CA50 SurrogateCalc CA50 PRF87

Pin kPa

CA

D

Pin kPa

T @

BD

C

47

LLNL-PRES-653562

Normalized HRR (PRF87)

Crank Angle relative to Peak HRR [ºCA]N

orm

aliz

ed H

RR

0

0.002

0.004

0.006

0.008

0.01

-35 -25 -15 -5 5

Hea

t-R

elea

se R

ate

/ To

tal H

R [1

/°C

A]

Crank Angle relative to Peak HRR [°CA]

Pin = 325 kPa

Pin = 240 kPa

Pin = 200 kPa

Pin = 180 kPa

Pin = 160 kPa

Pin = 130 kPa

Pin = 100 kPa

b.

PRF87 shows a much more intense early LTHR

Despite the very close ignition delay time profiles, the two fuels behave quite differently (Sensitivity Lack of NTC)

0

0.0002

0.0004

0.0006

0.0008

0.001

-35 -30 -25 -20 -15 -10 -5 0 5 10

HRR 100kPa NormHRR 130kPa NormHRR 160kPa NormHRR 180kPa NormHRR 200kPa NormHRR 240kPa NormHRR 324kPa Norm

Gasoline: Experiments PRF87: Calculations

48

LLNL-PRES-653562

Final Remarks

The evolution of engines and fuels proceeded in tandem since the dawn of the automotive industry with fuel quality being the limiting factor in engine development

In the last 90 years both engines and fuels have dramatically changed, but, for fuel properties specifications, we are still using fuel quality standards defined back then

Fuel surrogates allow us to run more reproducible experiments and sophisticated engine numerical analysis, though there are still many open questions on how a good surrogate should be formulated

Detailed kinetic models can be used to guide the formulation of surrogates and are effective tools to interpret fuel behavior in the combustion chamber

49

LLNL-PRES-653562

Premixed Air/Fuel

Mixture

Engine Combustion: Homogeneous Charge Compression Ignition (HCCI): two Stage Fuels

Typical HCCI Combustion Temperature and Heat Release Rate

profiles

Crank Angle Degree

Hea

t Rel

ease

Rat

e

Top Dead Center

1st Heat release stage

The HRR slows down

(Negative TemperatureCoefficient)

Main Heat Release

T, P

50

LLNL-PRES-653562

Typical HCCI Combustion Temperature and Heat Release Rate

profiles

H

•

R•

- RH

+ O2


OO•

OOH•

•OO

OOH

HOO

O

+ O2

- •OHO

O

•

•OH+ +

+ HO2•

+ •OH

+ •OHO

+

O

•

Fast High


T, P

CAD

HR

R

TDC

HCCI combustion kinetics: two Stage Fuels

Chemical Kinetics Modeling of Engine Processes

Charles Westbrook

Tsinghua-Princeton 2014 Summer School on Combustion

July 20-25, 2014

Kinetic analyses of engine processes• Stratified charge engine details• Flame quench in lean-burn engines• Flame quench on engine chamber walls• Heat transfer to engine walls during flame quenching• Flame quench in expanding engine chamber• Fuel additive studies• Detonation parameters• Engine knock and octane numbers, proknocks and antiknocks• Diesel ignition and cetane numbers• Biodiesel fuel combustion• Soot production• Reduction of soot in diesel engines with oxygen added• Pulse combustor optimization• HCCI combustion mechanisms• Surrogate fuel formulations• Reduced mechanisms for ignition and combustion• Ignition models

WSS meeting 10/2007 210/16/2007

Laminar flames in quenching problems

Mid-volume quenching in direct injection stratified charge (DISC) engine

Bulk quenching due to volume expansion in lean-burn engine mixtures

Flame quenching at lean and rich flammability limits

Flame quenching on cold walls and unburned hydrocarbon emissions from internal combustion engines

Emissions from DISC engine

• Interest in mid 1970’s by engine designers in stratified charge engine. Honda CVCC, Ford PROCO, GM DISC

• Principle was spatial charge stratification, achieved by spray injection.

• Flame should burn fuel, halt abruptly when it reached the discontinuity in fuel concentration.

• Experimental observations of Lancaster (GMR) of excessive UHC emissions.

WSS meeting 10/2007 610/16/2007

Flame quenching by volume expansion

• Concept of “lean burn” engine in mid-1970’s• Experimental studies in real engines by Quader

demonstrated large UHC emissions and flame failure for late-ignition, extremely lean fuel/air mixtures.

• “What limits lean operation in spark ignition engines - Flame initiation or propagation?”A.A. Quader, 1976.

• Laboratory experiments by Smith and Sawyer, with supporting laminar flame modeling, answered Quader’s questions.

WSS meeting 10/2007 710/16/2007

Expansion of thecombustion chamberreduces radical levelsbelow a critical value

WSS meeting 10/2007 810/16/2007

Flame quenching on engine walls

Previous concept of UHC emissions

Idea of making wall layers thinner

Simple flame model results at Ford weren’t believed

Detailed modeling results

Evidence had been there, Wentworth

WSS meeting 10/2007 910/16/2007

Pictures of wall quenching

Flame approaching wall Fuel diffuses away from walland is rapidly consumed

WSS meeting 10/2007 1010/16/2007

WSS meeting 10/2007 1110/16/2007

Heat transfer to walls during flame quench compares well with experimental results

Experiments (solid curves) from thin film resistance thermometer (Vosen, Greif and Westbrook, 1984)Calculations (dashed curves) using laminar flame with detailed chemical kinetics

WSS meeting 10/2007 1210/16/2007

WSS meeting 10/2007 1310/16/2007

Chemical Kinetics and Soot Production

Charles Westbrook

Tsinghua-Princeton 2014 Summer School on Combustion

July 20-25, 2014

2

Diesel engine combustion: A revolution

3

Early models of Diesel combustion

Liquid core with continuous evaporation (1976)

4

Early models of Diesel combustion

Liquid fuel jet shedding droplets, with combustion at the edge of a stoichiometric shell (diffusion flame)

Prior to Laser-Sheet Imaging

Autoignition and premixed burn were thought to occur in near-stoichiometric regions.

The "quasi-steady" portion of Diesel combustion was thought to be adequately described by steady spray combustion theory.

Appeared to fit most available data.

This "old" description was never fully developed into a conceptual model.A "representative" schematic is given.

Schematic of group combustion for a fuel spray.From Kuo, as adapted from H. Chiu and Croke

Old description of DI Diesel combustion.

The DOE Engine Combustion Research Program at Sandia’s CRF played a major role in solving the diesel “mystery”.

Mission - Develop the science-base for in-cylinder combustion and emissions processes.– Help U.S. manufacturers reduce

emissions & improve performance.

Approach –– Strong interaction and

collaboration with industry.– Optical diagnostics.– Realistic engine geometries with

optical access through:> pistons> cylinder liner> spacer plates> exhaust ports

Approach: Investigate the processes in the cylinder of an operating diesel engine using advanced optical diagnostics

Modified heavy-duty truck engine provides good optical access while maintaining the basic combustion characteristics of a production engine.

Data from multiple advanced laser diagnostics have substantially improved our understanding of diesel combustion and emissions formation.

Heavy-Duty Diesel Engine Research

Optical Setup

Laser-Sheet Imaging Data - 1Liquid-phase Fuel

Vapor-phase Fuel

Chemiluminescence

Liquid fuel images show that all the fuel vaporizes within a characteristic length (~1 inch) from the injector.

Vapor fuel images show that downstream of the liquid region, the fuel and air are uniformly mixed to an equivalence ratio of 3-4.

Chemiluminescence images show autoignition occurring across the downstream portion of the fuel jet.

Quiescent Chamber, 1200 rpm, TTDC = 1000 K, TDC = 16.6 kg/m3

Laser-Sheet Imaging Data - 2

PAH Distribution

Soot Distribution

OH PLIF Image

PAHs form throughout the cross-section of the fuel jet immediately following fuel breakdown at the start of the apparent heat release.

LII soot images show that soot forms throughout the cross-section of the fuel jet beginning just downstream of the liquid-fuel region.

OH radical images show that the diffusion flame forms at the jet periphery subsequent to an initial fuel-rich premixed combustion phase.

Quiescent Chamber, 1200 rpm, TTDC = 1000 K, TDC = 16.6 kg/m3

Laser Sheet Imaging is Providing aNew Understanding of DI Diesel Combustion

The appearance is significantly different.– Regimes of Diesel combustion are different than thought.

(flame standoff, upstream mixing, instantaneous vs. averaged).

Old Description New Conceptual Model

12

Predicting the soot precursors is one of the keys to predicting soot emissions from a Diesel engine

Fuel-Rich Premixed Reaction Zone

Fuel-rich premixed reaction zone

From: John Dec,SAE paper970873

Steady growthin molecular size leads to visible soot

Correlations between Fuel Structural Features and Benzene Formation

Hongzhi R. Zhang, Eric G. Eddings, Adel F. SarofimThe University of Utah

and Charles K. WestbrookLawrence Livermore National Lab

presented at2008 International Combustion Institute Meeting

Montreal, Canada, August 8th, 2008

Introduction

Chemistry of Benzene Precursors and Comparison of Measured and Predicted Benzene Concentrations

Benzene Formation Potential

Benzene Formation Pathways

Outline

IntroductionCombustion generated benzene is a health concern

Benzene is a major precursor for particulate pollutionBenzene is a known carcinogenBenzene is a major precursor of PAH, also carcinogens

We want to identify fuel properties that are critical to major benzene formation pathways and benzene formation potentials for individual fuel components

Fuel structure: between normal, iso-, and cyclo paraffinsFuel structure: C3, C4 vs. C12, C16 fuelsOther properties: Equivalence ratio, Hydrogen deficiency, Combustion temperature

22 premixed flames; C1-C12 fuels; = 1.0-3.06; P = 20-760 torr; Tmax = 1600-2370 K

Benzene concentrations were predicted within 30% of the experimental data for 15 flames (total of 22 flames)

Class 1: Acetylene addition (Westmoreland et al., 1989; Frenklach et al., 1985)

R1: C2H2 + CH2CHCHCH = C6H6 + HR2: C2H2 + HCCHCCH = C6H5

Class 2: C3 combination (Hopf, 1971; Miller-Melius, 1992)R3: H2CCCH + H2CCCH = C6H6R4: H2CCCH + CH2CCH2 = C6H6 + HR5: H2CCCH + H2CCCH = C6H5 + HR6: H2CCCH + CH2CHCH2 = FULVENE + 2H

Class 3: Combination of CH3 and C5H5R7: C5H5 + CH3 = C-C6H8 = C6H6 + 2H

Class 4: Cascading dehydrogenation (Zhang et al., 2007)R8: cycloC6-R C-C6H10(–R) C-C6H8(–R)C6H6(–R)

Class 5: De-alkylationR9: C6H5-R + H = C6H6 + R

Major Benzene Formation Pathways Revisited

Natural Gas,

Synfuel, Indicator

Fuels, Biofuels

Liquid Fuels

from Oil and Coal

Experimental: Benzene from Cyclo-Paraffins

Author Fuel P torr T(Max) in K [C6H6]V C7H16 1.0 760 1843 12 PPM HSP gasoline 1.0 760 1990 344 LWC C-C6H12 1.0 30 1960 473

Max

Questions:

1. Why gasoline produces more benzene than n-heptane, the indicator fuel for octane rating?2. What are the benzene sources in gasoline?3. How is benzene formed from various chemical classes?

Introduction




Outline

Sub-Models Compiled from Literatures

We took• Marinov-Westbrook-Pitz’s hydrogen model• Hwang, Miller et al.’s, and Westbrook’s acetylene oxid. models• Wang and Frenklach’s acetylene reaction set with vinylic and

aromatic radicals• Marinov and Malte’s ethylene oxidation sub-model• Tsang’s propane and propene chemical kinetics• Pitz and Westbrook’s n-butane sub-model• Miller and Melius benzene formation sub-model• Emdee-Brezinsky-Glassman’s toluene and benzene oxidation sub-

model

We have added• 100 modification steps to the base gas core concerning benzene

chemistry• Fuel Component Sub-Mechanisms

Precursor ChemistryA list of benzene precursors includes

Major precursors: C3H3, C2H2, n-C4H3, n-C4H5Minor precursors: C-C5H5, C-C6Hx, Ph-RBridging Species: a-C3H5, C2H3Other Related Species: a-C3H4, p-C3H4, C4H6 isomers, C3H5isomers, C-C5H6, C4H4, C4H2, C2-C4 olefins

New Reactions in the mechanismLarge olefin decomposition: 1-C7H14 = a-C3H5 + C4H9-1New addition of chemistry of p-C3H4

p-C3H4 has comparable, if not higher, concentrations in flames, in comparison with those of a-C3H4It is easier to form C3H3 radicals from p-C3H4 than from a-C3H4

Reactions involving C4 species

Reaction of C2H3=C2H2+H critically examined

BHC6H6

0.E+00

1.E-01

2.E-01

0 0.5 1HAB (cm)

X

TCBC6H6

0.E+00

2.E-02

4.E-02

6.E-02

0 0.2 0.4 0.6HAB (cm)

X

MPWCH4

0.E+00

1.E-04

2.E-04

3.E-04

0 0.3 0.6 0.9HAB (cm)

X

MPWC2H6

0.E+00

1.E-04

2.E-04

3.E-04

0 0.2 0.4 0.6 0.8HAB (cm)

X MCMC3H8

0.E+00

5.E-04

1.E-03

0 0.2 0.4 0.6 0.8HAB (cm)

X

CBLC4H6

0.E+00

1.E-03

2.E-03

0 0.5 1 1.5HAB (cm)

X

Modeled Benzene Concentrations

Modeled Benzene Concentrations

EDVC7H16

0.E+002.E-054.E-056.E-058.E-05

0 0.2 0.4 0.6HAB (cm)

X

EDVC8H18

0.E+00

2.E-04

4.E-04

6.E-04

0 0.2 0.4 0.6HAB (cm)

X

DDAC10H22

0.E+002.E-054.E-056.E-058.E-05

0 0.1 0.2 0.3HAB (cm)

X

HSP Gasoline

0.E+00

2.E-04

4.E-04

6.E-04

0 0.05 0.1HAB (cm)

X

DDAKerosene

0.E+00

1.E-03

2.E-03

0 0.1 0.2 0.3HAB (cm)

XLWC

C6H12

0.E+00

2.E-04

4.E-04

6.E-04

0 0.1 0.2 0.3HAB (cm)

X

Introduction




Outline


X means “a Factor of X”

# Fuel Inert Ar, % C/O Eq. Ratio

P torr T(Max) K at cm

Exp. Max. [C6H6]b at cm

Cal. Max. [C6H6]b at cm

Deviation

F1 CH4 0.453 0.626 2.50 760 1605 at 0.4 280 at 0.8 141 at 0.8 -49.6 F2 C2H6 0.453 0.715 2.50 760 1600 at 0.24 230 at 0.8 205 at 0.8 -10.9 F3 C3H8 0.44 0.833 2.78 760 1640 at 0.4 840 at 0.35 922 at 0.32 +9.8 F4 C3H8 0.424 0.54 1.80 30 2190 at 0.95 17.5 at 0.75 72.9 at 0.77 +4.2e

F5 C2H2 0.05 0.959 2.40 20 1901 at 1.0 40 at 0.37 82.7 at 0.37 +2.1e F6 C2H2 0.45 1.00 2.50 19.5 1850 at 1.0 58.9 at 0.6 39.1 at 0.64 -33.6 F7 C2H2 0.55 1.103 2.76 90 1988 at 0.73 140 at 0.6 96.7 at 0.55 -30.9 F8 C2H4 0.5 0.634 1.90 20 2192 at 1.7 33.1 at 0.9 11.4 at 0.77 -2.9e F9 C2H4 0 0.80 2.40 760 1815 at 0.1 936 at 0.15 136 at 0.14 -6.9e F10 C2H4 0.656 0.92 2.76 760 1600 at 0.3 250 at 0.35 212 at 0.35 -15.2 F11 C2H4 0.578 1.02 3.06 760 1420 at 0.3 575 at 1.0 553 at 1.0 -3.8 F12 C3H6 0.25 0.773 2.32 37.5 2371 at 0.71 1220 at 0.39 927 at 0.39 -24.0 F13 C4H6 0.03 0.874 2.40 20 2310 at 1.65 1300 at 0.85 1490 at 0.85 +14.6 F14 C6H6 0.3 0.717 1.79 20 1905 at 0.2 N/A N/A Good F15 C6H6 0.752c 0.72 1.80 760 1850 at 0.45 N/A N/A Good F16 C7H16 0.841c 0.318 1.00 760 1843 at 0.25 12 at 0.08 1.77 at 0.09 -6.8e F17 C7H16 0.73c 0.605 1.90 760 1640 at 0.30 75 at 0.225 75.8 at 0.23 +1.1 F18 i-C8H18 0.682c 0.608 1.90 760 1670 at 0.30 292 at 0.21 455 at 0.23 +55.8 F19 C10H22 0.682c 0.558 1.73 760 1688 at 0.20 65 at 0.10 68.5 at 0.10 +5.4 F20 gasoline 0.768c, 0.01d 0.9-1 760 1990 at 0.046 344 at 0.05 330 at 0.05 -4.1 F21 kerosene 0.684c 1.7 760 1775 at 0.20 1090 at 0.1 850 at 0.75 -22.0 F22 C-C6H12 0.325 0.333 1.00 30 1960 at 0.6 473 at 0.09 498 at 0.09 +5.3

Benzene Formation PotentialThe Highest and Lowest Benzene Producer

Author Fuel Inert Ar, %

C/O P torr

Exp. Max. Y(C6H6)b

Cal. Max. Y(C6H6)b

Deviation, %

CBL C4H6 0.03 0.874 20 1300 at 0.85 1490 at 0.85 +14.6 AHB C3H6 0.25 0.773 37.5 1220 at 0.39 927 at 0.39 -24.0 DDA kerosene 0.684c =1.7 760 1090 at 0.1 850 at 0.75 -22.0 CDB C2H4 0 0.80 760 936 at 0.15 136 at 0.14 -X6.9e MCM C3H8 0.44 0.833 760 840 at 0.35 922 at 0.32 +9.8 EDA C7H16 0.73c 0.605 760 75 at 0.225 75.8 at 0.23 +1.1 DDA C10H22 0.682c 0.558 760 65 at 0.10 68.5 at 0.10 +5.4 BDR C2H2 0.45 1.00 19.5 58.9 at 0.6 39.1 at 0.64 -33.6 WHL C2H2 0.05 0.959 20 40 at 0.37 82.7 at 0.37 +X2.1e BW C2H4 0.5 0.634 20 33.1 at 0.9 11.4 at 0.77 -X2.9e CNT C3H8 0.424 0.54 30 17.5 at 0.75 72.9 at 0.77 +X4.2e

V C7H16 0.841c 0.318 760 12 at 0.08 1.77 at 0.09 -X6.8e

Benzene Formation PotentialEffects of Carbon Backbone: C3 Species

Fuel decompositionC3H8 C3H6 a-C3H5 C3H4 C3H3

C3H8 C2H4 C2H2 C3H3

Benzene formationC3H3 + C3H3 = bC6H6

C3H3 + C3H3 = C6H5 + HC3H3 + a-C3H4 = bC6H6 + HC3H3 + a-C3H5 = fC6H6 + 2H

Author Fuel C/O P torr

T(Max) K at cm

T(Max), K at cm, Fitted

Exp. Max. Y(C6H6)b

Cal. Max. Y(C6H6)b

Deviation, %

AHB C3H6 0.773 37.5 2371 at 0.71 1220 at 0.39 927 at 0.39 -24.0 HW C2H4 0.92 760 1600 at 0.3 250 at 0.35 212 at 0.35 -15.2 CMM C2H4 1.02 760 1420 at 0.3 575 at 1.0 553 at 1.0 -3.8 MCM C3H8 0.833 760 1640 at 0.4 840 at 0.35 922 at 0.32 +9.8 MPW CH4 0.626 760 1605 at 0.4 280 at 0.8 141 at 0.8 -49.6 MPW C2H6 0.715 760 1600 at 0.24 230 at 0.8 205 at 0.8 -10.9

Benzene Formation PotentialEffects of Carbon Backbone: C4 Species

Fuel decompositionC4H6 C4H5 C4H4 C4H3

C4H6 C2H3

C4H5 C2H3 & C2H2

C4H5 + H C3H3Benzene formation

C2H2 + C4H3 = C6H5

C2H2 + C4H5 = bC6H6 + HC3H3 + C3H3 = bC6H6


T(Max) K at cm


Exp. Max. Y(C6H6)b

Cal. Max. Y(C6H6)b

Deviation, %

CBL C4H6 0.874 20 2310 at 1.65 2050 at 1.75 1300 at 0.85 1490 at 0.85 +14.6 WHL C2H2 0.959 20 1901 at 1.0 40 at 0.37 82.7 at 0.37 +X2.1e

BDR C2H2 1.00 19.5 1850 at 1.0 58.9 at 0.6 39.1 at 0.64 -33.6

Benzene Formation PotentialEffects of Carbon Backbone: Cyclohexanes

Benzene formationCascading dehydrogenation & Interweaving dehydrogenation

C-C6H12 C-C6H10 C-C6H8 bC6H6

R-C-C6H11 C-C6H10 C-C6H8 bC6H6

R-C-C6H11 R-C-C6H9 C-C6H8 bC6H6

R-C-C6H11 R-C-C6H9 R-C-C6H7 bC6H6

R-C-C6H11 R-C-C6H9 R-C-C6H7 R-C6H5


T(Max) K at cm


Exp. Max. Y(C6H6)b

Cal. Max. Y(C6H6)b

Deviation, %

DDA kerosene =1.7 760 1775 at 0.20 1775 at 0.25 1090 at 0.1 850 at 0.75 -22.0 DDA C10H22 0.558 760 1688 at 0.20 1688 at 0.22 65 at 0.10 68.5 at 0.10 +5.4 LWC C-C6H12 0.333 30 1960 at 0.6 1960 at 0.55 473 at 0.09 498 at 0.09 +5.3 V C7H16 0.318 760 1843 at 0.25 12 at 0.08 1.77 at 0.09 -X6.8e

HSP gasoline =1 760 1990 at 0.046 1990 at 0.106 344 at 0.05 330 at 0.05 -4.1

Benzene Formation PotentialEffects of Branching: cyclo > iso > normal paraffins

Fuel decompositioni-C8H18 i-C4H8 i-C4H7

i-C4H7 a-C3H4 C3H3

i-C4H8 s-C3H5 p-C3H4 C3H3Benzene formation

C3H3 + C3H3 = bC6H6

C3H3 + C3H3 = C6H5 + HC3H3 + a-C3H4 = bC6H6 + H


T(Max) K at cm


Exp. Max. Y(C6H6)b

Cal. Max. Y(C6H6)b

Devion, %

EDA i-C8H18 0.608 760 1670 at 0.30 1670 at 0.36 292 at 0.21 455 at 0.23 +55.8EDA C7H16 0.605 760 1640 at 0.30 1640 at 0.40 75 at 0.225 75.8 at 0.23 +1.1DDA C10H22 0.558 760 1688 at 0.20 1688 at 0.22 65 at 0.10 68.5 at 0.10 +5.4 LWC C-C6H12 0.333 30 1960 at 0.6 1960 at 0.55 473 at 0.09 498 at 0.09 +5.3V C7H16 0.318 760 1843 at 0.25 12 at 0.08 1.77 at 0.09 -X6.8

Introduction




Outline


1.0810-7

C6H5CH3

C6H5OH2CCCH+H2CCCH

5.11 10-7

8.4610-7

C6H6

2.0310-7

Fulvene

1.3110-7

H2CCCH+CH2CCH2

1.2710-7

C3Hx+C4Hx

1.25 1

0-6

C6H5CH21.7810-7

2.3810-7

1.0310-7H2CCCH+ CH2CHCH2

H2CCCCH+C2H2

C6H51.2910-6

1.08 1

0-7

5.3710-9

C2Hx+C4Hx

1.2410-7

Rates

Magnitude

Precursors

In a Normal Decane Flame

1.0810-7

C6H5CH3

C6H5OH2CCCH+H2CCCH

5.11 10-7

8.4610-7

C6H6

2.0310-7

Fulvene

1.3110-7

H2CCCH+CH2CCH2

1.2710-7

C3Hx+C4Hx

1.25 1

0-6

C6H5CH21.7810-7

2.3810-7


H2CCCCH+C2H2

C6H51.2910-6

1.08 1

0-7

5.3710-9

C2Hx+ C4Hx

1.2410-7

Contribution of Major Benzene Formation Pathways51% from C3H3 + C3H3 = bC6H6

13% from C3H3 + a-C3H5 = fC6H6 + 2H13% from C3H3 + a-C3H4 = bC6H6 + H12% from C2H3 + C4H3 (C4H5) = C6H5 (bC6H6 + H)11% from C6H5-CH3 + H = C6H6 + CH3

In an Acetylene Flame


5% from C6H5-CH3 + H = C6H6 + CH3Propargyl Radical Formation Pathways

87% 1CH2 + C2H2 = H2CCCH + H11% 3CH2 + C2H2 = H2CCCH + H

1.1210-9

C6H5CH3

C6H5O

H2CCCCH+C2H2

C6H6

4.510-9

Fulvene

8.3010-11

H2CCCH+CH2CCH2

8.4710-11

C3Hx+C4Hx

1.34 1

0-9

1.0310-9

C6H5CH2


C6H5

5.9210-9

1.3210-8

1.36 1

0-9

H2CCCH+H2CCCH 2.4610-8

1.96 10-8

In a Cyclohexane Flame

Contribution of Major Benzene Formation Pathways100% from cycloC6-R C-C6H10 C-C6H8 bC6H6

7.1210-9C6H5CH3

C3Hx+ C3H3

9.66 1

0-9

1.0010-8

C6H6

2.7210-7

C4Hx +C2H2

5.2610-10

C-C6H10

1.52 1

0-6

C6H5

2.49 1

0-8

C-C6H11

3.1910-6

C-C6H12 2.8010-5

C-C6H8

1.4410-6

Benzene Formation Pathways: in Butadiene Flames


20% from C2H2 + C4H3 (C4H5) = C6H5 (bC6H6 + H)12% from C3H3 + a-C3H5 = fC6H6 + 2H10% from C6H5-CHO + H = C6H6 + CHO5% from C6H5-CH3 + H = C6H6 + CH3

5% from C5H5 + CH3 = C-C6H8 = C6H6 + 2H

CH2CHCHCH+C2H2 1.0510-7 3.4210-8

C6H5CH3

H2CCCH+H2CCCH

3.29 10-7

6.1210-7

H2CCCCH+C2H2

C6H6

3.2510-8

Fulvene

8.4110-8

C3Hx+C4Hx

1.31 10-6

H2CCCH+ CH2CHCH2

7.4110-8

C6H5CHO

6.6210-8

C6H5CH2

2.4110-8

2.3510-7

7.3310-7

C6H5CO

4.32 1

0-8

C6H5C2H

C6H5O6.4410-7

C5H5

3.54 1

0-8

C-C6H7

3.5410-8

C6H5

7.63 1

0-8

4.4910-7

8.7110-8

6.1510-7

CH2CHCCH2+C2H2

3.7910-8

C6H5CHO4.2110-7

7.6310-8

48% from C3H3 + C3H3 = bC6H6

20% from C2H2 + C4H3 (C4H5) = C6H5 (bC6H6 + H)

Soot Precursor Production Potential

Contribution from Individual Surrogate Components to the Formation of Benzene (a kerosene fuel)

Benzene Contributors: Benzene (28%), Toluene (26%) and Methyl Cyclohexane (40%)Component Fractions: Benzene (1%), Toluene (10%), Methyl Cyclohexane (10%), and paraffins (79%)

n-C12H26

a

i-C8H18

C6H11CH3

C6H5CH3C6H6

n-C12H26

b

i-C8H18

C6H11CH3

C6H5CH3

C6H6

Surrogate Distribution Benzene Formation

* Experimental: Doute et al., Combustion Science Technology, 106 (4-6) (1995) 327–344.* Modeling: Zhang et al., Proceedings of Combustion Institute, (2007) 31, 401-409.

Concluding CommentsThe Utah Surrogate Model Was Validated for 22 Premixed Flames of Various Fuels (C1-C12 fuels; = 1.0-3.06; P = 20-760 torr; T = 1600-2370 K).Benzene Concentrations Were Predicted within 30% of the Experimental Data for 15 (out of 22) Flames.Both Formation Pathways and Formation Potential of Benzene Were Found to Be Dependent on the Fuel Structure, and C3, C4 and C-C6 Were among the Most Productive Fuels.C3 Combination Was Identified to be the Major Benzene Formation Pathway for Most Fuels; That Is Replaced with Dehydrogenation Only for Cyclohexanes.Acetylene Addition Was Found to Be Important in C4Flames and Those with Large Paraffinic Fuels.

Flame experiments on formation of the first aromatics, usually benzene or toluene McNesby et al., Combust. Flame 142, 413-427 (2005)

Opposed flow diffusion flames

McEnally et al., Prog. En. Comb. Sci. 32, 247-294 (2006) Co-flow non-premixed flames

Violi and Izvekov, PROCI 31, 529-537 (2007). Molecular dynamics

Desgroux et al., PROCI 34, 1713-1738 (2013) Review of use of optical diagnostics

Thomas and Wornat, PROCI 32, 2009. Pyrolysis of catechol and butadiene at 1000 C, residence time of 0.3 seconds

Premixed ignition in Diesel combustion

• Fuel-rich conditions ( ≈ 4 )• Relatively low temperature (T ≈ 850 K )

- Source of cetane ratings in Diesel engines- Very similar to conditions of engine knock

- Very complex chemical kinetic pathways

• Products are good producers of soot precursor species

• Ignition kinetics are the same as in engine knock in SI engines, driven by H2O2 decomposition

Products of rich premixed ignition are mostlysmall unsaturated hydrocarbons, especially acetylene and ethene,which are known precursors to soot

Experimental background

• Addition of oxygenated species reduces soot

- Important possible oxygenates include biodiesel fuels

• Soot production correlates with post-ignition levels of selected

chemical species

• Suggestions that this is due to presence of C - C bonds or total O

concentrations

• Use kinetic model to examine these possibilities

44

Predicted level of soot precursors correlates well with soot emissions from a Diesel engine

From:Flynn, Durrett, Dec, Westbrook, et al., SAE paper1999-01-0509

45

Structure of Tripropylene Glycol Monomethyl Ether (TPGME)

Experiments at Sandia show same trends as LLNL kinetic models

DBM and TPGME reduce sooting, but DBM is less effective than TPGME

Models show same soot precursor formation, but oxygen enhances precursor consumption

Example of C2H3 + O2 breaking C - C bond

Understand and predict emissions from open burning or detonation of explosives

RDX and HMX are based on non-aromatic rings

C C NN N N C

C C C NN N C

Note the absence of C - C bonds or aromatic rings

NO2

NH2NO2

NH2

O2N

H2N

CH3NO2

NO2

O2N

TATBTNT

Presence of aromatic rings indicates explosivewill lead to soot.

Aromatic rings and lots of C - C bonds

RDX does not produce soot precursors

N

NNNN

N

O

OO

O

OO

N

N.N

N

N

O

O

OO

+ NO2

CH2N

N

O

O

CH2N

N

O

O

CH2

N.

+

+

+ NO2CH2

.N

No carbon – carbon bonds!

From Ree et al, J. Phys. Chem. A, 1996

Soot tendencies depend on molecular structure

Oxygen is the main obstacle to soot production

• Goal is to produce C - O bonds

• There are many possible sources of oxygen

- Simplest alternative is air

- Oxygenated hydrocarbon or other molecules

• This is the principle used in diesel engines to reduce soot production

• This is the explanation for some munitions combustion observations

Molecular structure of oxygenated fuel additive determines its soot reduction properties Variability in soot precursor production observed computationally

Before modeling approach was used, all oxygenates were believed to

be equally effective at soot reduction

Subsequent engine experiments consistent with model results

Reaction pathways that lead to early CO2 production “waste” available

oxygen atoms in the oxygenate

Same approach provided sooting estimates for oil sands fuel

All analysis based on single-component “diesel fuel” surrogate

Need for more thorough, multicomponent diesel simulations

Opportunities for designing optimal oxygenated additives

63

HCCIHigherO2

Low temperatureTemperature [K]

Kitamura, et al. 2002

ConventionalDiesel path

Documents

Detailed Kinetic Modeling of Gasoline Fuels