HCCI and Stratified-Charge CI Engine Combustion Research

May 19, 2009 – 10:00 a.m.

U.S. DOE, Office of Vehicle TechnologiesAnnual Merit Review and Peer Evaluation

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

Program Manager: Gurpreet Singh Project ID: ace_04_dec

John E. DecWontae Hwang and Yi Yang

Sandia National Laboratories

2

OverviewTimeline

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

• Project directions and continuation are evaluated annually.

Budget• Project funded by DOE/VT:

FY08 – $695kFY09 – $700k

Barriers• Extend HCCI (LTC) operating range

to higher loads.• Improved understanding of in-cylinder

processes.• Control HC & CO emiss. at low loads.

Partners / Collaborators• Project Lead: Sandia ⇒ John E. Dec• Part of Advanced Engine Combustion

(AEC) working group:– 15 Industrial partners: auto, engine & energy– 5 National Labs & Univ. of Wisconsin

• GM – bimonthly meetings & discussion• Chevron – funds complementary project• LLNL – 1) support kinetic-mechanism

devel., 2) CFD modeling, & 3) cooperative project on detailed exhaust speciation.

3

Objectives

FY09 Objectives:• Determine the development of natural thermal stratification in an HCCI

engine, using planar-imaging thermometry.– Thermal-imaging diagnostic developed as part of this task.

• Evaluate the potential of intake boost for extending the high-load limit of HCCI by using EGR to control combst.-phasing advance – multi-year task.– FY09: Determine potential of boost with EGR for gasoline at rep. engine speed.

• Determine the performance of ethanol as a fuel for HCCI engines.– Conducted cooperatively with M. Sjöberg in the Advanced SI-Engine Fuels Lab.

• Support CFD modeling and the development/improvement of chemical-kinetic mechanisms for HCCI at LLNL ⇒ provide data and analysis.

Project objective: to provide the fundamental understanding (science-base) required to overcome the technical barriers to the development of practical HCCI and HCCI-like engines by industry.

4

MilestonesFY2008

• Complete analysis of detailed exhaust-gas speciation measurements for iso-octane. (February 2008) – Status: Completed

• Determine the potential benefits of EGR for reducing the maximum pressure-rise rate and extending the high-load HCCI limit. (August 2008) – Status: Completed

FY2009

• Determine the magnitude and distribution of the natural thermal stratification in an HCCI engine at a typical operating condition. (February 2009) – Status: Completed.

• Determine the potential of EGR for increasing the allowable intake-pressure boost for gasoline-like fuels, at a representative engine speed. (August 2009) – Status: ~60% complete as of March 2009.

5

Approach

• Metal engine ⇒ design well-characterized experiments to isolate specific aspects of HCCI/SCCI combustion & relationships between parameters.– Intake boost: Systematically increase boost ⇒ adjust Tin and/or EGR to retard

timing to allow max. fueling at each Pin without knock, but with good stability.

• Optical engine ⇒ detailed investigations of in-cylinder processes.– Thermal stratification (TS): Develop temperature-imaging diagnostic ⇒ Apply to

obtain T-map images showing temporal and spatial development of TS.

• Computational Modeling ⇒ supplement experiments by showing cause-and-effect relationships that are not easily measured. – Initiating LES modeling with J. Oefelein, Sandia to understand mechanism of TS.– In-house CHEMKIN (Senkin) single- and multi-zone kinetic modeling.– Collaborate with LLNL to improve kinetic mechanisms, and on CFD modeling.

• Combination of techniques provides a more complete understanding.

• Transfer results to industry.

• Use a combination of metal- and optical-engine experiments and modeling to build a comprehensive understanding of HCCI processes.

6

All-Metal Engine

Optical Engine

Optics Table

Dynamometer

Intake Plenum

Exhaust Plenum

Water & Oil Pumps & Heaters

Flame Arrestor

Sandia HCCI / SCCI Engine Laboratory

• Matching all-metal & optical HCCI research engines.– Single-cylinder conversion from Cummins B-series diesel.

Optical EngineAll-Metal Engine

• Bore x Stroke = 102 x 120 mm • 0.98 liters, CR=14

7

Accomplishments• Determined the evolution of natural thermal stratification in an HCCI engine,

including its distribution and magnitude at a typical operating condition. – Developed a planar temp.-imaging diagnostic for TS in HCCI engines.

• Conducted initial investigation showing the potential of intake boost for extending the high-load limit of HCCI for gasoline fuel.– Showed that EGR is effective for controlling boost-induced timing advance.– Achieved a substantial load increase at a rep. 1200 rpm operating condition.

• Determined the behavior of ethanol as an HCCI fuel over a range of operating conditions.– Cooperatively with M. Sjöberg of the Advanced SI-Engine Fuels Lab.

• Initiated detailed exhaust-speciation analysis for PRF80 ⇒ 2-stage ignition.– Project conducted in cooperation L. Davisson at LLNL.

• Supported chemical-kinetic and CFD modeling work at LLNL.– Provided data and analysis for: 1) improving chemical-kinetic mechanisms, and

2) CFD modeling of fuel stratification to improve low-load comb. eff. & emissions.

8

Importance of Thermal Stratification (TS)• TS causes autoignition to occur

sequentially from hottest region to coldest.– Reduces max. pressure-rise rate (PRR).– Allows higher fueling without knock.

• Amplify the benefit of the TS by retarding combst. timing ⇒ further increases in load.

• Chemilum. images show:– Non-uniformities over whole field of view.– Hot reactions start intermittently near the

mid-plane.– At time of max. PRR most combustion is

from bulk gases (central region).– BL combust. occurs after max. PRR.

• TS of the bulk gas is critical for high-load HCCI operation.

• Understanding TS is important for increasing the high-load limit of HCCI.

20

25

30

35

40

45

50

55

60

360 364 368 372 376Crank Angle [° ATDC]

Cyl

inde

r Pre

ssur

e [b

ar]

0

30

60

90

120

150

180

210

240

HR

R [J

/°CA

] & P

RR

x 1

0 [b

ar/°C

A]

PressureHRRPRR * 10 Chemilum.

Side-ViewBottom-View

9

Planar Imaging Thermometry• Diagnostic: single-laser toluene PLIF.

– PLIF intensity varies with temperature.– Good sensitivity in desired range, 600 – 1050 K.

• PLIF setup:– 2% toluene + 98% iso-octane– Laser excitation: 266 nm, 58 mm wide sheet.– Intensified camera with 277nm LP & UG5 filters.– Run inert with N2 to prevent quenching.⇒ OK since TS develops prior to combustion.

• Calibrate temp. sensitivity in-cylinder.

• For well-mixed fueling, variations in PLIF intensity correspond to temp. variation– Temperature fluctuations shown relative

to the mean of each image.– Bright regions in raw image correspond to

cold pockets.

Nd:YAG266 nm

RawPLIF

T-map

10

• Laser elevation adjusted with crank angle to remain in mid-plane (20 - 4 mm below F-D).– Representative of bulk gas.

• TS develops progressively as cold pockets convected into central region.– Temperature nearly uniform at 305°CA.

> Virtually no TS remains from intake.> Insufficient time with Tgas > Twall.

– Substantial TS by TDC (360°CA) +/- 35 K.> Sufficient for significant spread in

autoignition time of various regions.

• TS distribution is random cycle-to-cycle.

• Scale of cold pockets near TDC is 5–11 mm, similar to 8mm TDC clearance height.(Fine-grain speckle pattern is shot noise.)

• Magnitude of TS appears to diminish after TDC.

Temporal Evolution of TS

340°CA305°CA 330°CA

360°CA350°CA 355°CA

365°CA 370°CA 380°CA

390°CA

TDC

11

0

10

20

30

40

50

60

300 320 340 360 380 400Crank Angle [° CA]

Ther

mal

Wid

th [K

]

0.01

0.02

0.03

0.04

0.05

0.06

0.07

TW/T

MEA

N

TW - 5 to 95%TW/Tmean

• Apply probability density functions (PDF) to quantify changes in temp. distribution with crank angle.– 305°CA: PDF very narrow.

> Little time for development of TS.> Analysis ⇒ almost all width is shot noise.

– 330 - 340°CA: Significant broadening.– 340 - 360°CA: Progressive increase.– 360 - 390°CA: PDF width decreases, in

agreement with images. ⇒ cause?

• Define: thermal width (TW) = 5 -95% of PDF width.– Max. TW at TDC ≈ 50 K ⇒ agrees with

multi-zone model results.

• Normalize by TMEAN to remove effects of compression & expansion.– Reduces, but not eliminate TW decrs. ATDC.

PDF Analysis of TS Images

0.00

0.03

0.06

0.09

0.12

0.15

-40 -30 -20 -10 0 10 20 30 40ΔT [K]

Mas

s PD

F

305

320

330

340

345

350

355

360

0.00

0.03

0.06

-40 -30 -20 -10 0 10 20 30 40ΔT [K]

Mas

s PD

F 360365370375380390

Before TDC

After TDC

• Heat transfer dominant ⇒ changes at TDC.

12

Thermal Distribution of Boundary Layer (BL)• Incrementally scan laser from mid-plane

to outer BL (to 0.8 mm below firedeck).

• 330°CA: bulk-gas temp. is nearly uniform. ⇒ Significant TS only for z/h ≥ -0.43.

• 360°CA: TS developed throughout bulk gas. ⇒ TS greater in outer BL, z/h ≥ -0.5.

• Avg. T profiles also show deficits for these z/h.

mid-plane 330°CA

z/h = -1 z/h = -0.76 z/h = -0.54

z/h = -0.22 z/h = -0.09z/h = -0.43

360°CA

z/h = -1 z/h = -0.75 z/h = -0.5

z/h = -0.25 z/h = -0.2z/h = -0.38

mid-plane

750

800

850

900

950

1000

1050

-1.0 -0.8 -0.6 -0.4 -0.2 0.0z/h

Ave

rage

Tem

pera

ture

[K]

360 CAD330 CAD

• TS progresses inward from wall, 330°→360°• Most BL temp. deficit occurs in last 0.8 mm

at the wall ⇒ drops to Twall ≈ 400 K.

• BL thickness based on a 5% deficit from centerline value ⇒ 1-1.5mm.

• Agrees with previous chemilum. & PLIF studies.

13

0

2

4

6

8

10

12

14

16

18

50 100 150 200 250 300 350Intake Pressure [kPa abs.]

IMEP

g [b

ar]

Pin = 320 kPaPin = 280 kPaPin = 240 kPaPin = 200 kPaPin = 160 kPaPin = 100 kPa

0

2

4

6

8

10

12

14

16

18

50 100 150 200 250 300 350Intake Pressure [kPa abs.]

IMEP

g [b

ar]

Pin = 320 kPaPin = 280 kPaPin = 240 kPaPin = 200 kPaPin = 160 kPaPin = 100 kPa

0

2

4

6

8

10

12

14

16

18

50 100 150 200 250 300 350Intake Pressure [kPa abs.]

IMEP

g [b

ar]

Pin = 320 kPaPin = 280 kPaPin = 240 kPaPin = 200 kPaPin = 160 kPaPin = 100 kPaMax. IMEPg

• Investigate the potential of boosting for extending HCCI to higher-loads. – Required to match full-load diesel or SI.

• Current work: gasoline, 1200 rpm.

• Boost enhances autoignition ⇒advances comb. timing ⇒ Knock!– Compensate with reduced Tin.– For Pin > 160 kPa, Tin → Tamb⇒ limits allowable fueling.

• Add cooled EGR to further slow autoignition.

Intake Boost for Extending High-Load Limit

Increase Fueling

0

1

2

3

4

5

6

50 100 150 200 250 300 350Intake Pressure [kPa abs.]

Rin

ging

Inte

n. [M

W/m

2]

& C

orr.

Std-

Dev

. of I

MEP

g [%

]

Ringing IntensityCorr. Std-Dev. of IMEPg

At Max. IMEPg• Achieved IMEPg = 16.3 bar, Pin = 324 kPa. – Very high IMEPg for HCCI/LTC, convent’l fuel.– Near stoich., C/F = 38.5, EGR = 60%,

Pexhaust = 326 kPa, Texhaust = 407°C.

• Ringing ≤ 5 MW/m2, No Knocking.

• Std-Dev of IMEPg ≤ 1%, very good stability.

14

10001100120013001400150016001700180019002000

50 100 150 200 250 300 350Intake Pressure [kPa abs.]

Peak

Cha

rge

Tem

pera

ture

[K]

0.00.10.20.30.40.50.60.70.80.91.0

NO

x [g

/kg-

fuel

]

Peak Charge Temp.NOx

At Max. IMEPg

• NOx emissions below US-2010 stds.(should also meet tier II, bin 5).– Extremely low for all boosted cases

( < 0.1 g/kg-fuel, ~1-2 ppm).

• Correlates with low peak charge temp.– NOx higher for Pin = 100, Tpeak > 1900K.

For maximum IMEPg at each boost• Indicated Thermal Eff. increases

slightly with boost. ⇒ Th. Eff. ~45%.

• Combustion Eff. increases, 97→ 99%.– Higher wall temps. ⇒ improve combst.– Increased EGR reduces HC & CO emiss.

Efficiency and NOx for Boosted High-Load

10

15

20

25

30

35

40

45

50

50 100 150 200 250 300 350Intake Pressure [kPa abs.]

Indi

cate

d Th

erm

al E

ff. [%

]

90

92

94

96

98

100

102

104

106

Com

bust

ion

Effic

ienc

y [%

]

Thermal Eff.Combustion Eff.

At Max. IMEPg

• For max. IMEPg =16.3 bar, Pin =3.2 bar.– Ind. Thermal Eff. = 47%– Comb. Eff. = 99%– NOx = 0.015 g/kg-fuel, Tpeak = 1750 K.

15

• Ethanol is a component in most pump gasoline (0 – 15% fraction).– Also being considered as an alternative fuel at levels up to 85%.

• Important to understand ethanol’s potential for HCCI.– Ignition quality ⇒ RON = 107, MON = 89– Effect on performance and operating

range, i.e. speeds, boost, etc.

Ethanol as an HCCI Fuel

20

40

60

80

100

120

140

160

180

300 600 900 1200 1500 1800 2100 2400Engine Speed [rpm]

Inta

ke T

empe

ratu

re [°

C]

PRF90 EthanolPRF80 GasolinePRF70 30%nC7, 40% Tol, 30% iC8

φ = 0.38CA50 = 372°CAPin = 100 kPa

• Speed sweep ⇒ CA50 = 372°CA– Autoig. similar to gasoline for RPM >900.– Most fuels: Tin < 100°C as speed is

reduced ⇒ indicates LTHR (cool flame).– Ethanol shows no LTHR!

• Boost has only moderate effect on Tin.

100

120

140

160

180

90 100 110 120 130 140 150Intake Pressure [kPa]

BD

C T

empe

ratu

re [°

C]

Ethanol, ModelEthanol, Experiment

φ = 0.38CA50 = 372°CA

1500 rpm

• Performs similarly to gasoline, conds. studied. • Good potential for HCCI fuel / fuel-component.• Ethanol is a true single-stage ignition fuel.

– May offer advantages for control & boosted oper.

16

Detailed Exhaust Speciation – PRF80• Joint project with LLNL ⇒ spec. analysis.

Sandia ⇒ engine op. & data interpretation.

• Conducted fueling(φ)-sweep & data for near-misfire conditions. Results provide:– Data for aftertreatment & model validation.– Improved understanding of combst. process.

• PRF80 is a 2-stage ignition fuel at conditions studied. ⇒ Affects emissions compared to iso-octane & gasoline.– OHC fraction is greater for PRF80.– Unreacted-fuel fraction is much lower. ⇒ Cool-flame reactions incr. fuel breakdown.

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36Equivalence Ratio [φ]

Perc

ent F

uel C

arbo

n [%

]

HC + OHCHC (no OHC Carbon)OHC (all OHC Carbon)Fuel = Iso-Oct + n-hep.Iso-OctaneEtheneMethane2-Methyl-1-PropeneAcetone2-Methyl-2-Propenaln-Heptane

Main Species

0.00.10.20.30.40.50.60.70.80.91.0

0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36Equivalence Ratio [φ]

Perc

ent F

uel C

arbo

n [%

]

EtheneMethane2-Methyl-1-PropeneAcetone2-Methyl-2-PropenalFormaldehydeIsomers, Dimethyl-Pentenen-HeptanePhenoln-Heptene2,4,4-Trimethyl-1-pentene

Selected Species

• Ratio of n-heptane / iso-octane is 19-23%, compared to 25% in fuel.– n-Heptane breaks down more readily, but it

induces substantial iso-octane breakdown.

• Relatively high conc. of n-Heptene & Phenol ⇒ former due to n-Heptane reactions.

17

Future Work• Complete investigation of intake boost for extending the high-load limit of

gasoline-fueled HCCI at a representative speed,1200 rpm (FY09).

• (FY10) Expand boost study to include a range of higher engine speeds, boost levels, and back-pressures for realistic turbo-charger efficiencies.– Two-stage fuels to be done as part of Chevron-funded project.

• Extend TS study: 1) improve diagnostic S/N & optical setup, 2) investigate methods of increasing TS, and 3) determine cause of flows producing TS. – Collaborate with J. Oefelein to apply LES modeling ⇒ mechanism / enhancement.

• Additional ethanol studies over a wide range of operating parameters: EGR, load, & boost to high levels ⇒ with M. Sjöberg, Adv. SI-Fuels Lab.

• Complete exhaust-speciation analysis for 2-stage ignition fuel, PRF80, and compare with single-stage ignition fuels, gasoline and iso-oct. ⇒ with LLNL.– Analyze emiss. species for near misfire with single- and two-stage ignition fuels.

• Continue to collaborate with LLNL on improving chemical-kinetic mechanisms and on CFD/kinetic modeling.

18

Summary• Quantitative temperature-map images show that thermal strat. (TS) develops

progressively during latter compression stroke ⇒ throughout charge.

• Data indicate that TS results from wall-heat transfer and convection.– Future work will focus on understanding the mechanism for bulk-gas TS, and

potential methods for increasing it to increase the high-load limit of HCCI.

• EGR substantially improves boosted HCCI operation with gasoline fuel.– Achieved 16.3 bar IMEPg, Ind. Thermal-Eff. = 47%, no Knock & no NOx or PM. – Near high-load limit for conventional diesel. Shows significant potential for

extending HCCI range – full time HCCI?

• Ethanol is a promising HCCI fuel. Performance is generally similar to gasoline, but no low-temp. (“cool-flame”) chemistry. – Possible advantages for control and for boosted op. ⇒ additional studies req’d.

• Detailed exhaust speciation of a two-stage ig. fuel (PRF80), shows significantly different behavior from single-stage fuels, iso-octane & gasoline.– More breakdown of fuel & fuel-like species to smaller species, higher OHC fract.