Low-Temperature Diesel Combustion Cross-Cut Research

Lyle M. PickettSandia National Laboratories

Sponsor: DOE Office of Vehicle TechnologiesProgram Managers: Gurpreet Singh and Kevin Stork

FY 2010 DOE Vehicle Technologies Program Annual Merit ReviewProject ACE005, V Virginia C, 11:00 – 11:30 AM, Tuesday, June 8, 2010

This presentation does not contain any proprietary, confidential, or otherwise restricted information.

Project ID: ACE005

2/25

Overview

• Project provides fundamental research that supports DOE/ industry advanced engine development projects.

• Project directions and continuation are evaluated annually.

• Engine efficiency and emissions

– Sources of unburned hydrocarbons and CO for LTC combustion

• Load limitations for LTC• CFD model improvement for

engine design/optimization

• Project funded by DOE/VT:FY09- $570KFY10 - $660K

Timeline

Budget

Barriers

Partners• 15 Industry partners in the

Advanced Engine Combustion MOU

• Participants in the Engine Combustion Network– Experimental and modeling

• Project lead: Sandia – Lyle Pickett (PI)

3/25

Overall Approach

• Facility dedicated to fundamental combustion research for both heavy-duty and light-duty engines (cross-cut research).

– Well-defined charge-gas conditions• Pressure, temperature, EGR level

– Well-defined injector parameters• Injection pressure, fuel, multi-injections

Experiments in CV•Well-defined boundary

conditions•Quantitative diagnostics at

engine conditions• Improved physical

understanding

High-Efficiency, Low-Emissions Engine

Computer models• “Retain” insights from experiment• Adds knowledge about things that

are not “measurable”• Parametric design optimization• Saves time and cost over “hardware”

iteration

4/25

Objectives/Milestones

• Aid the development of computational models for engine design and optimization (ongoing).

– Experimental and modeling collaboration through the Engine Combustion Network: http://www.ca.sandia.gov/ECN

– FY10 (1): Attain injector set for experimentation by multiple laboratories. Characterize “Spray A” high-temperature, high-pressure condition.

• Determine the factors that cause liquid wall impingement at post-injection (DPF-regeneration) conditions (FY09-FY10).

– Urgent need to understand oil dilution, inefficiency using diesel/biodiesel fuel.– FY10 (2): First vaporization measurements at low-density, high-T conditions.

• Quantitative characterization of soot processes in transient sprays at LTC conditions (FY08-FY10).

– Emitted soot is an emissions burden, source of inefficiency, expense.– FY10 (3): Comparison of biodiesel and diesel soot distributions. Methods to

reduce soot even with mixing-controlled (negative-dwell) heat release. – FY10 (4): Measurement of soot size distribution using thermophoretic sampling.

Critical information needed for soot modeling at LTC conditions.

5/25

(1) Industrial and academic collaboration in the Engine Combustion Network

Michigan Tech. Univ.Vessel temperature composition

Argonne (x-ray source)Internal needle movementNear-nozzle liquid volume

IFP (France)Spray velocityCombustion

CMT (Spain)Rate of injectionSpray momentum

SandiaLiquid and vapor mixingCombustion diagnostics

Meiji (Japan)Soot formation

Internal nozzle geometry

CaterpillarLift-off lengthLiquid length

General MotorsMixingDroplet sizingBosch

Donates 10 “identical” common-rail injectors

• Operation at the same ambient and injector conditions: “Spray A”

6/25

Experimental participation in the ECNInstitution Facilities PersonnelSandia Preburn CV Lyle Pickett, Caroline GenzaleIFP Preburn CV Gilles Bruneaux, Louis-Marie MalbecCMT Cold CV,

Flow PVJulien Manin, Raul Payri

Chalmers Flow PV Mark LinneGM Flow PV Scott ParrishMich. Tech. U. Preburn CV Jeff Naber, Jaclyn Nesbitt, Chris MorganArgonne Cold V, X-ray

SynchrotronChris Powell, Alan Kastengren

Caterpillar Flow PV Tim Bazyn, Glen MartinAachen Flow PV Heinz PitschMeiji U. Preburn CV Tetsuya AizawaSeoul Nat. U. Preburn CV Kyoungdoug Min

7/25

Past/future modeling participation in the ECNInstitution LeadSandia Joe OefeleinIFP Christian AngelbergerChalmers Fabian KärrholmPolit. di Milano Tommaso LucchiniCity Univ. Manolis GavaisesImperial College David GosmanUniv. Wisconsin-Madison Rolf Reitz, Chris RutlandDoshisha Univ. Jiro SendaAachen/Stanford Heinz PitschPurdue Univ. John AbrahamPenn State Dan HaworthConvergent Science Kelly Senecal

Many other companies (like Bosch) and institutions…..

8/25

Ambient gas temperature 900 KAmbient gas pressure 6 MPaAmbient gas density 22.8 kg/m3Ambient gas composition 15% O2, 75.1% N2, 6.2% CO2, 3.6% H2OCommon rail fuel injector Bosch solenoid-activated, generation 2.2Fuel injector nozzle outlet diameter 0.090 mmNozzle K factor 1.5 { K = (dinlet – doutlet)/10 [use µm] }Nozzle hydro-erosion Discharge coefficient = 0.86 with 100 bar ∆P.Spray full included angle 0° (1 axial hole) or 145° (3-hole )Fuel injection pressure 150 MPaFuel n-dodecaneFuel temperature at nozzle 363 K (90°C)Common rail volume/length 22 cm3/ 28 cm (Use GM rail model 97303659)Distance, injector inlet to common rail 24 cmFuel pressure measurement 7 cm from injector inlet / 24 cm from nozzleInjection duration 1.5 msApproximate injector driver current 18 A for 0.45 ms ramp, 12 A for 0.345 ms hold

How will focus at “Spray A” conditions advance understanding of spray combustion?

• Quantitative, complete datasets (velocity, species, temperature) still do not exist at engine conditions.

– Need to move beyond “conceptual model” to more detailed comparison. – Follows direction of successful activities using more basic flames (TNF workshop).

• Leveraging of work by many experimental and modeling activities will accelerate research.

-15 -10 -5 0 5 10 150

0.05

0.1

0.15

0.2

0.25

Radius [mm]

Mix

ture

Fra

ctio

n

x = 30 mm ExperimentModel 1Model 2Model 3Model 4Model 5

Spray A Specifications Model comparisons to baseline dataset available on the ECN

9/25

Careful characterization of vessel boundary conditions (temperature) is required.

Gas temperature measurement

Schlieren of preburn and cooldown

500 1000 1500 2000 2500 3000400

600

800

1000

1200

1400

1600

1800

2000

Time after spark [ms]

Tem

pera

ture

[K]

Tbulk

ZRMWPT

ubulk ⋅⋅

⋅=ρ

Compression Heating

Axial x = 55 mmHoriz. y = 0 mmVertical z = 0 mm

Flame arrival

Vessel position for TC

10/25

Ambient gas has uniform, horizontal “core”.Core gas temperature distribution

0 10 20 30 40 50 60 700.94

0.96

0.98

1

1.02

1.04

1.06

Axial distance from injector nozzle [mm]

T cor

e,m

easu

red/

T cor

e,pr

edic

ted

TC1, y = -15, z = 0TC2, y = 0, z = 0TC3, y = 0, z = -15TC4, y = 0, z = +15TC5, y = +15, z = 0

Buoyancy induced vertical variation in temperature

11/25

Injector boundary conditions characterized at multiple facilities.

18 Amps 450 µs +12 Amps 345 µs

Spray A specification: 1500 µs injection durationmass = 3.5 mg (here)

330 µs hydraulic delay

n-dodecane90 C1 atm back pressure

0 500 1000 1500 20000

1

2

3

4

5M

ass

flow

rate

[mg/

ms]

Time ASI [µs]

110

120

130

140

150

160

Fuel

Pre

ssur

e 7

cm fr

om in

ject

or [M

Pa]

Pressure

Current/10

Mass flow

IFP

IFP

Sandia

Sandia

Sandia 26

cm fr

om n

ozzl

e

12/25

0 500 1000 1500 2000 2500 3000 35000

10

20

30

40

50

60

70

80

90

100

Max

imum

Pen

etra

tion

[mm

]

Time ASI [µs]

SandiaIFP

Liquid

Vapor Non-reacting

Reacting Penetration

Comparison of vapor/liquid penetration at two facilities shows reasonable agreement.Schlieren visualization.Liquid Mie-scatter border in blue.0% O2: non-reacting

San

dia

IFP

mov

iem

ovie

13/25

Ignition at 450 µs is followed by a quasi-steady lift-off length at 15-17 mm.

0 500 1000 1500 2000 2500 3000 3500 40000

10

20

30

40

50

Lift-

off l

engt

h [m

m]

Time after start of injection [µs]

< 450 nm chemiluminescence (Sandia)

1500 µs duration

“steady” lift-offOH chemiluminescence (IFP)Time-average 1100-1800 µs

Distance from injector [mm]0 5 10 15 20 25 30 35 40 45 50 55 60 65 70

-15

-10

-5

0

5

10

15

• Ignition/lift-off are examples of other measurements for Spray A.• Refined, quantitative measurements are being pursued to follow these

initial measurements (e.g., mixture fraction, liquid volume fraction).

mov

ie

14/25

(2) Liquid penetration for late-cycle post-injections is problematic, uncharacterized.• One strategy used for DPF regeneration is to implement a late-cycle

post-injection to enrich the exhaust stream.

900

950

1000

1050

1100

1150

1200

1250

1300

1350

1400

1450

60 75 90 105 120 135 150

Tem

pera

ture

[K]

CA° ATDC

3.0kg/m3

2.1kg/m3

1.6kg/m3

1.4kg/m3 1.2

kg/m3

TBDC = 313 KPBDC = 100 kPaA/F = 20:1CR = 17.5:1

2. Relatively high temperatures (post-combustion)

1. Very low densities

Typical post-injection timing for DPF regeneration

HOWEVER,• These injections typically occur well after TDC (near EVO), when in-

cylinder densities are quite low. Wall-wetting, oil dilution are concerns!

15/25

800 1000 1200 140020

40

60

80

100

120

Temperature [K]

Qua

si-S

tead

y Li

quid

Len

gth

[mm

]

800 1000 1200 140020

40

60

80

100

120

Temperature [K]

Qua

si-S

tead

y Li

quid

Len

gth

[mm

]

Even though temperature are high, liquid penetration exceeds engine dimensions.

1

3 kg/m3

2 kg/m3 1.2 kg/m3 3 kg/m3

2 kg/m31.2 kg/m3

Neat Soy-based Biodiesel:15-25% increase in liquid penetration#2 Diesel

½ bore: light-duty engines

Biodiesel351

793

Fuel #2 DieselT90 [°C] 315

ρ at 100°C [kg/m3] 767

0.108 mm nozzle, 1500 bar injection pressure

16/25

0 0.2 0.4 0.6 0.8 10

20

40

60

80

100

Time ASI [ms]

Max

imum

Liq

uid

Pen

etra

tion

[mm

]

0 0.2 0.4 0.6 0.8 10

20

40

60

80

100

Time ASI [ms]

Max

imum

Liq

uid

Pen

etra

tion

[mm

]Short injections can limit the maximum liquid penetration, but fuel quantities are small

1

0.5 mg

0.65 mg

0.5 mg

long injection

0.65 mg

1.2 kg/m3800 K

1.2 kg/m3800 K

3.0 kg/m31200 K

mov

iem

ovie

17/25

(3) First measurements of soot distribution within combusting biodiesel sprays.

• Factor of five reduction in total soot formed within reacting spray.• Long inception time permits soot-free combustion with negative

ignition dwell, lower heat release.

1

D2

BD

Distance from Nozzle [mm]0 10 20 30 40 50 60 70 80

51015

Soot volume fraction [ppm]

#2 Diesel

Soy Methyl EsterBiodiesel

1000 K67 bar15% O2

Flame Lift-off

18/25

TEM probe within the combustion vessel

Spray directionTEM grid

TEM grid held within the probe to protect it from excessive heating

(4) Soot sampling from sprays within high-temperature, high-pressure combustion vessel.

3 mm diameter TEM grid:

copper mesh covered by an amorphous carbon film

0 100 200 300 400 500

0

100

200

300

400

500

size

[nm

]

Soot-laden gas skims past the probe, enters through a 1-mm diameter hole, and deposits on the TEM grid.

Jet

Restrictive passage quenches reaction and protects the TEM grid.

TEM image of soot particles

0 100 200 300 400 500

0

100

200

300

400

5001

20 30 40 50 60 70 80

5

10

15

20

size

[nm

]

30 mm collection position

20 100 200 300 400 500

0

100

200

300

400

500

20 30 40 50 60 70 80

5

10

15

20

size

[nm

]

50 mm collection position

2

20 30 40 50 60 70 80

5

10

15

20

0 100 200 300 400 500

0

100

200

300

400

500

size

[nm

]

60 mm collection position

0 100 200 300 400 500

0

100

200

300

400

5002

20 30 40 50 60 70 80

5

10

15

20

size

[nm

]

86 mm collection position

23/25

50 mm 86 mm

50 mm 86 mm

A key input for soot models, the particle size and shape are now characterized.

23% m-xylene77% n-dodecane

0 10 20 30 40 50 60 70 80

5

10

15

20

0 100 200 300 400 500

0

100

200

300

400

5000 100 200 300 400 500

0

0

0

0

0

0

Peak soot location

Soot oxidation region

Size [nm]

24/25

Future work

• Outlook for the Engine Combustion Network– Measurements (not a complete list, Sandia in blue):

– Side-hole spray compared to axial hole– Liquid volume fraction near the liquid length

• Spray-spray interaction effects on mixing and lift-off length.• Biodiesel soot processes, including soot particle morphology.• Gasoline direct injection sprays

– Isolate ignition processes using laser ignition.– Identify regimes with both autoignition and flame propagation.– Use diesel and GDI-type injectors.

• Nozzle shape • Internal needle movement• Discharge and area

contraction coefficients• Rate of injection• Near-nozzle liquid volume

fraction

• Droplet size, velocity, shape• Maximum liquid penetration• Vapor penetration rate• Velocity and turbulence

within spray• Mixture fraction (non-

reacting and reacting)

• Ignition delay• Cool flame position and

timing• Heat-release rate• Quantitative soot, soot

precursor distribution• Lift-off length

25/25

Presentation Summary

• Project is relevant to the development of high-efficiency, low-emission engines.

– Observations of combustion in controlled environment lead to improved understanding/models for engine development.

• FY10 approach addresses long term and short term needs.– Lead team to develop the ECN Spray A condition.– Massive dataset is being generated, which will be a key component for future

model improvement.– Responding to new problems for DPF regeneration, characterized late-cycle

post-injection liquid penetration.– Made first measurements of soot distribution in biodiesel sprays.– Quantified the soot size and morphology within sprays at engine conditions.

• Collaboration expanded to accelerate research and provide greatest impact (MOU, Engine Combustion Network)

• Future plans will continue effort– “Spray A” characterization for the ECN.

Low-Temperature Diesel Combustion Cross-Cut ResearchOverviewOverall ApproachObjectives/Milestones�(1) Industrial and academic collaboration in the Engine Combustion NetworkExperimental participation in the ECNPast/future modeling participation in the ECN�How will focus at “Spray A” conditions advance understanding of spray combustion?Careful characterization of vessel boundary conditions (temperature) is required.Ambient gas has uniform, horizontal “core”.Injector boundary conditions characterized at multiple facilities.Comparison of vapor/liquid penetration at two facilities shows reasonable agreement.Ignition at 450 ms is followed by a quasi-steady lift-off length at 15-17 mm.(2) Liquid penetration for late-cycle post-injections is problematic, uncharacterized.Even though temperature are high, liquid penetration exceeds engine dimensions.Short injections can limit the maximum liquid penetration, but fuel quantities are small(3) First measurements of soot distribution within combusting biodiesel sprays.(4) Soot sampling from sprays within high-temperature, high-pressure combustion vessel.Slide Number 19Slide Number 20Slide Number 21Slide Number 22A key input for soot models, the particle size and shape are now characterized.Future workPresentation SummaryAdditional slides: Response to previous year’s reviewer comments:PublicationsCritical assumptions and issuesOptical diagnostics experimental setup