HCCI and Stratified-Charge CI Engine Combustion … and Stratified-Charge CI Engine Combustion Research May 10, 2011 – 9:30 a.m. U.S. DOE, Office of Vehicle Technologies Annual Merit

HCCI and Stratified-Charge CI Engine Combustion Research

May 10, 2011 – 9:30 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: ACE004

John E. DecNicolas Dronniou and Yi YangSandia 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:

FY10 – $750kFY11 – $750k

Barriers• Extend HCCI (LTC) operating

range to higher loads.• Increase the efficiency of HCCI

(LTC).• Improve the understanding of

in-cylinder processes.

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

working group – 15 industrial partners• General Motors – specific collaboration• LLNL – support kinetic modeling• Univ. of Michigan• Univ. of New South Wales, Australia• Chevron• LDRD – advanced biofuels project

(internal Sandia funding)

3

Objectives - Relevance

FY11 Objectives ⇒ High Loads, Increased Efficiency, Improved Understanding

• Thermal Stratification (TS): Determine 1) the main sources of colder near-wall gases, and 2) the primary mechanisms for transport and dispersion of this colder gas into the hotter bulk gas, at a base operating condition.– Initiate investigation of how operating conditions affect the development of TS.– Improve PLIF-based thermal imaging technique for side-view imaging.

• Improved Efficiency of Boosted HCCI: Examine various operating techni-ques to determine their potential for increasing the efficiency of intake-boosted HCCI (e.g. effects of Tin, CA50, ringing, DI vs. pre-mixed fueling).

• Continue collaborations with J. Oefelein (Sandia) to conduct LES modeling to better understand the mechanisms producing TS & how to improve TS.

• Support chemical-kinetic and CFD modeling of HCCI at LLNL, the Univ. of Michigan and General Motors ⇒ 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 or HCCI-like engines by industry.

4

Approach

• Metal engine ⇒ conduct well-characterized experiments to isolate specific aspects of HCCI/SCCI combustion.– Improved efficiency: Select representative boost, determine effect of parameters

of interest (Tin, CA50, fueling method) while holding other key parameters const.

• Optical engine ⇒ detailed investigations of in-cylinder processes.– Thermal stratification: Apply PLIF-based thermal-imaging using a vertical laser

sheet to simultaneously image both the boundary layer (BL) and bulk gas.

• Computational Modeling ⇒ supplement experiments by showing cause-and-effect relationships that are not easily measured. Also, to improve models. – Collaborate w/ J. Oefelein (Sandia) on LES modeling to understand mech. of TS.– Support LLNL & U of Mich. to improve kinetic mechanisms & on CFD modeling.

• Combination of techniques provides a more complete understanding.

• Transfer results to industry: 1) physical understanding, 2) improved models, 3) data to GM to support their in-house modeling of TS & boosted HCCI.

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

5

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 Engine

All-Metal Engine

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

Metal-engine experiments ⇒ Fuel is gasoline: RON = 91.7, MON = 83.4

6

Accomplishments• Determined the main sources of colder near-wall gases and the mechanism

for dispersing this colder gas into the hotter bulk-gas to produce TS.– Significantly improved side-view thermal imaging technique.– Preliminary data for effect of operating parameters on development of TS.

• Evaluated effects of fueling method on efficiency: premixed, DI, & Partial-DI.– Showed significant efficiency improvements with DI & partial-DI for boosted HCCI.

►Showed that partial-DI fueling allows a substantial increase in the high-load limit of boosted HCCI ⇒ gasoline becomes φ-sensitive with boost.

• Evaluated the effects of intake temperature (Tin), combustion timing (CA50), and ringing-intensity on engine efficiency.

► Investigated benefits of partial-DI fueling with ethanol, collab. with M. Sjöberg.

• Collaborating with J. Oefelein on LES modeling to supplement TS-imaging experiments ⇒ developed new high-fidelity grid, computations underway.

• Supported chemical-kinetic and CFD modeling work at LLNL, the Univ. of Michigan and General Motors ⇒ provided data and analysis.

7

• Increasing TS has strong potential for extending the high-load limit of HCCI.– And/or increasing efficiency.

• A better understanding of the mechanism(s) producing TS is needed.

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

sequentially from hottest region to coldest.– Reduces max. pressure-rise rate (PRR) ⇒ allows higher loads and better efficiency.

• At time of max. PRR most combustion is from bulk gases (central region).

340°CA

370°CA

TDC - 360°CA

355°CA

T-Map Imagesmid-plane

Bottom-View

1368°CA 1368°CA 1368°CA

Side-View

372° 10

1

CA of Max. PRR368°

12364°

1.4366°

372° 10372° 10

1

CA of Max. PRR368° 1

CA of Max. PRR368°

12364° 12364°

1.4366° 1.4366°

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

6

6

6

6

6

6

6

6

6

6

7

0

10

20

30

40

50

-10

-20

-30

-40

-50

∆T [K]

6

6

6

6

6

6

6

6

6

6

7

0

10

20

30

40

50

-10

-20

-30

-40

-50

∆T [K]

0102030405060708090

100

340 350 360 370 380 390 400 410Crank Angle [°CA ATDC]

Cyl

inde

r Pre

ssur

e [b

ar] Model, Phi = 1.0

Exp., Phi = 0.44

• Prev. T-map images show TS development in bulk gas, late in compression stroke.

8

Simultaneous Bulk-Gas and BL Imaging• T-maps from PLIF with toluene tracer.

– Excite with Nd:YAG @ 266 nm.– Run inert with N2 to prevent quenching.– FY10 ⇒ motored or fired gives same TS.– Calibrate in-situ by varying Tin.

• Optical setup allows visualization of bulk-gas and boundary-layer regions.– View extends to cylinder wall because

window acts as a divergent lens.

• Image post-processing into T-map significantly more challenging.– Repeatable cold BL & fluctuating temp.– Vignetting and beam steering.

End view of laser sheet

Raw ImageTemperature map (T-map)

9

Cycle-Averaged T-maps

Compression Stroke

• Boundary layers develop at cyl. walls and firedeck.– Compression suppresses

thickness of BL at firedeck.

• Avg. temperature is fairly uniform throughout bulk-gas.

Expansion Stroke

• BL expansion and crevice out-gassing increase TS.– Not relevant to controlling

max. PRR.

• Average T-maps show TS that is consistent from cycle to cycle.– Boundary layers (BL) and out-gassing during early expansion.

• There is no evidence of consistent flows that transport cold gas from near-wall regions into central bulk gas.

10

Single-Cycle T-maps

• Substantial TS is evident in the bulk gas near TDC.

• Bulk-gas TS is dominated by a randomly fluctuating pattern of cold pockets.

• Random nature suggests it results from in-cylinder turbulence.

• Characterized by turbulent structures of cold gas extending from the walls.– Not isolated cold pockets.

• Bulk-gas TS ⇒ critical for controlling max. PRR ⇒ appears to result from in-cylinder turbulence ⇒ turbulent structures extending from the wall.

• Show all TS ⇒ 1) random fluctuations, and 2) consistent from cycle to cycle.– Selected images show typical amounts of TS.

11

Analysis of Cold-Pocket Structures

• T-maps converted to binary images of cold pockets.

• Pockets categorized as:– attached firedeck– attached to piston top – unattached, in bulk-gas

• Number of cold pockets and their fraction of the total image area increase up through TDC.

• Less cold area at piston top may be related to high T of quartz piston top.

• Most cold pockets are structures attached to firedeck or piston top.– “Unattached bulk-gas” pockets may

be attached out of image plane.

• Statistical analysis of turbulent cold pockets shows size and location.

InitialT-map

BinaryImage

IdentifiedCold Pockets

FluctuatingComponent

360 CA

12

Total thermal stratification

Fluctuating thermal stratification

Average thermalstratification

Magnitude of TS• Std-Dev. of T-map provides quantitative measure of TS.

– Corrected for shot noise (<1.5 K).

• Total TS increases continuously.– Average TS increases due to BL development & out-gassing after TDC.

• Fluctuating component of TS reaches a maximum at TDC.– In agreement with SAE 2009-01-0650.

• Max. std-dev at TDC ≈ 16 K ⇒ Agrees with values required for multi-zone modeling of metal-engine burn duration (SAE 2005-01-0113).

13

43.544.044.545.045.546.046.547.047.548.0

20 30 40 50 60 70 80 90 100Intake Temperature [°C]

Indi

cate

d Th

erm

al E

ff. [%

]

DI, CA50 = 376.7

PM, CA50 = 376.7

Early DI vs. PreMixedFueling = 55 mg/inj

43.544.044.545.045.546.046.547.047.548.0


Indi

cate

d Th

erm

al E

ff. [%

]

DI, Ringing = 5DI, CA50 = 376.7PM, Ringing = 5PM, CA50 = 376.7

Early DI vs. PreMixedFueling = 55 mg/inj

370

371

372

373

374

375

376

377

378


CA

50 [°

CA

]DI, CA50 = 376.7

PM, CA50 = 376.7

Improving Efficiency of Boosted HCCI• FY10: Showed Thermal-Eff. (T-E) could

increase from 43.5 → 47.5% by adjusting various operating parameters.

• FY11: Conduct systematic sweeps to determine mechanisms and trade-offs.

• Example: Vary Tin with constant fueling.1. Const. CA50 ⇒ T-E up as Tin reduced.

> Higher γ and less heat-transfer loss.> No advantage of DI fueling over PM.

2. Const. Ringing = 5 MW/m2 (const. PRR).⇒ PM fueling: T-E similar to const. CA50 since is ringing similar. ⇒ Early-DI fueling: T-E much higher.

> DI reduces HRR, so can advance CA50 to get ringing = 5 ⇒ higher T-E.

Pin = 2 bar, Fuel = 55 mg/inj, Gasoline

370

371

372

373

374

375

376

377

378


CA

50 [°

CA

]DI, Ringing = 5DI, CA50 = 376.7PM, Ringing = 5PM, CA50 = 376.7• Why does early-DI give better performance?

– PLIF images show incomplete mixing ⇒ mixture stratification.

14

363

364

365

366

367

368

369

370

371

0.3 0.32 0.34 0.36 0.38 0.4 0.42 0.44 0.46 0.48 0.5Charge Mass Equivalence Ratio [φm]

CA

10 [°

CA

]

368

369

370

371

372

373

374

375

376

CA

10 [°

CA

]

PRF73, Pin = 1 bar, base-phi = 0.42

Gasoline, Pin = 1 bar, base-phi = 0.42

Pin varies

363

364

365

366

367

368

369

370

371


CA

10 [°

CA

]

368

369

370

371

372

373

374

375

376

CA

10 [°

CA

]

PRF73, Pin = 1 bar, base-phi = 0.42



Pin varies

363

364

365

366

367

368

369

370

371


CA

10 [°

CA

]

368

369

370

371

372

373

374

375

376

CA

10 [°

CA

]

PRF73, Pin = 1 bar, base-phi = 0.42Gasoline, Pin = 1 bar, base-phi = 0.42Gasoline, Pin = 2 bar, base-phi = 0.345Gasoline, Pin = 2 bar, base-phi = 0.42

Pin varies

363

364

365

366

367

368

369

370

371


CA

10 [°

CA

]

368

369

370

371

372

373

374

375

376

CA

10 [°

CA

]

PRF73, Pin = 1 bar, base-phi = 0.42Gasoline, Pin = 1 bar, base-phi = 0.42Gasoline, Pin = 1.6 bar, base-phi = 0.367Gasoline, Pin = 2 bar, base-phi = 0.345Gasoline, Pin = 2 bar, base-phi = 0.42

Pin varies

Φ-Sensitivity of Gasoline

• Use Fire19/1 technique to isolate fuel-chemistry effects from thermal effects.– Dec & Sjöberg SAE 2004-01-0557.

• Sweep φ above & below base fueling.

• Pin = 1 bar ⇒ chemistry not φ-sensitive; like iso-octane, γ effect dominates.

• Pin = 2 bar ⇒ strong φ-sensitivity more than PRF73.

• Pin = 1.6 bar ⇒ intermed. φ-sensitivity.

• With boost, gasoline autoignition strongly φ-sensitive. Not for Pin = 1 bar.

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21Cycle #

Pres

sure

φ = variable

Aquired cycle

φ = 0.20Fire 19/1

• For mixture stratification to reduce HRR, fuel autoignition must be sensitive to variations in the local φ.– Prev. thought to require a 2-stage

ignition fuel (e.g. PRF73).

15

• Partial fuel stratification (PFS) ⇒ premix most fuel & late DI up to 20%.– Vary DI timing or DI% to vary stratification.

• Large drop in ringing with increased DI%.

• Increased DI% ⇒ more regions of higher φm⇒ autoignite faster ⇒ advances hot ignition for same CA50 ⇒ increases burn duration.– Reduces peak HRR, PRRmax, and Pmax.

• Ultra-low NOx & soot. COV of IMEPg < 1.5%.

Pin = 2 bar: Controlled Mixture Stratification

50

60

70

80

90

100

110

340 350 360 370 380 390Crank Angle [°CA]

Pres

sure

[bar

]

3% DI @ 300°CA6% DI @ 300°CA9% DI @ 300°CA13% DI @ 300°CA17% DI @ 300°CA20% DI @ 300°CA

-5

0

5

10

15

340 350 360 370 380 390Crank Angle [°CA]

PRR

[bar

/°CA

]


-1000

100200300400500600700

340 350 360 370 380 390Crank Angle [°CA]

HR

R [J

/°CA

]


Hold: φm = 0.44, Tin = 60 C, CA50 = 374 CA

0

2

4

6

8

10

12

14

0 5 10 15 20 25DI Fuel [%]

Rin

ging

Inte

nsity

[MW

/m2 ]

10

11

12

13

14

15

16

17In

take

O2 C

onc.

[%]

Ringing IntensityIntake O2

16

5

6

7

8

9

10

11

12

13

14

0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6Charge-Mass Equiv. Ratio [φm]

IMEP

g [b

ar]

364

368

372

376

380

384

388

392

396

400

CA

50 [°

CA

]

IMEPg, Fully PreMixed

CA50, PreMixed

5

6

7

8

9

10

11

12

13

14


IMEP

g [b

ar]

364

368

372

376

380

384

388

392

396

400

CA

50 [°

CA

]

IMEPg, Full DI

IMEPg, Fully PreMixed

CA50, Full DI

CA50, PreMixed

Extend High-Load Limit with PFS, Pin = 2 bar

• PFS allows a large increase in load for gasoline boosted to Pin = 2 bar.

• Reduced ringing with PFS allows higher fueling and/or advanced CA50. • Premixed fueling ⇒ increase load from φm = 0.3 to knock/stability limit.

– Retard CA50 so Ringing ≤ 5 MW/m2 ⇒ Limit: IMEPg = 11.7 bar at φm = 0.47.

5

6

7

8

9

10

11

12

13

14


IMEP

g [b

ar]

364

368

372

376

380

384

388

392

396

400

CA

50 [°

CA

]

IMEPg, Full DIIMEPg, Partial DI (PFS)IMEPg, Fully PreMixedCA50, Full DICA50, Partial DICA50, PreMixed

• Full early-DI fueling:– Some mixture stratification– Allows advanced CA50, but stability

limit is IMEPg = 10.7 bar, φm = 0.40.

• PFS (partial DI) ⇒ much better stability ⇒ allows signif. higher loads.– Limit: IMEPg = 13.0 bar, φm = 0.54.– Approaching oxygen availability limit

⇒ 0.9% O2 in exhaust.

• Large reduction of HRR (PRR) allows – CA50 more advanced than premixed.– Higher thermal eff. and ringing reduced to 2 – 3 MW/m2 ⇒ greater stability.

17

Load-Limit & Eff. Improvements at Various Pin

• Our initial investigation of boosted HCCI in SAE 2010-01-1086 showed ⇒ maximum load attainable for well premixed HCCI at various boost levels.

8

9

10

11

12

13

14

15

16

1.4 1.6 1.8 2 2.2 2.4 2.6 2.8Intake Pressure [bar]

Max

imum

IMEP

g [b

ar]

0

1

2

3

4

5

CO

V of

IMEP

g [%

]

IMEP, PM, SAE 2010-01-1086

COV_IMEP, PM 2010-01-1086

8

9

10

11

12

13

14

15

16

1.4 1.6 1.8 2 2.2 2.4 2.6 2.8Intake Pressure [bar]

Max

imum

IMEP

g [b

ar]

0

1

2

3

4

5

CO

V of

IMEP

g [%

]

IMEP, PM, SAE 2010-01-1086IMEP, Partial DI (PFS)COV_IMEP, PM 2010-01-1086COV_IMEP, Partial DI (PFS)

43.0

43.5

44.0

44.5

45.0

45.5

46.0

46.5

47.0

892

883

1057

1082

1105

1094

1128

1192

1249

1295

1318

1335

1397

1410

1434

Average IMEPg [kPa] at various Pin [bar]

Indi

cate

d Th

erm

al E

ff. [%

] Premixed Partial Stratification

Pin = 1.8Pin = 1.6 Pin = 2.2Pin = 2.0 Pin = 2.4

R ≈ 2 R ≈ 4.5

IMEPg

[kPa]

O2 Limited• PFS allows higher loads for all Pin tested.– PFS quite stable Pin ≥ 2.0 bar

⇒ largest gain.– O2 limited for Pin ≥ 2.2 bar.– PFS less stable for Pin = 1.6 & 1.8 bar.

> Lower φ-sensitivity.

• Alternatively, PFS can allow CA50 advance for same ringing (PRR).

• T-E increases from 0.3 to 1.6% ⇒fuel economy gain of 0.7 – 3.6%.

• PFS gives typical fuel economy improvement of 2 – 2.5%.

18

Ethanol-Fueled HCCI – Effects of PFS

363364365366367368369370371372

0.35 0.37 0.39 0.41 0.43 0.45 0.47 0.49Charge Mass Equivalence Ratio [φm]

CA

10 [°

CA

]

368369370371372373374375376377

CA

10 [°

CA

]

PRF73, Pin = 1 barGasoline, Pin = 1 barEthanol, Pin = 1 barGasoline, Pin = 2 bar

2530354045505560

354 356 358 360 362 364 366 368 370 372 374Crank Angle [°CA]

Pres

sure

[bar

]

SOI = 40°CA, 20% DISOI = 280°CA, 20% DISOI = 280°CA, 30% DISOI = 280°CA, 40% DI

-50

0

50

100

150

200

250

300

354 356 358 360 362 364 366 368 370 372 374Crank Angle [°CA]

Hea

t-Rel

ease

Rat

e [J

/°CA

]

SOI = 40°CA, 20% DISOI = 280°CA, 20% DISOI = 280°CA, 30% DISOI = 280°CA, 40% DI

• PFS reduces HRR & PRR with ethanol, butbenefit is much less than for boosted gasoline.

• Ethanol is an important alternative fuel.• Exhibits true single-stage ignition

– Autoignition chemistry not φ-sensitive.– Very temperature sensitive.

• Also, a high heat of vaporization & much lower γ for a given φ than gasoline.

• Combination results in an inverse φ-sensitivity compared to boosted gasoline.

• Can we exploit this to reduce the HRR by using PFS to increase the TS?

• PRR and HRR are reduced with PFS due to increased TS. – Ignites lean-to-rich, opposite of boosted

gasoline with PFS.– NOx just below US-2010 at conditions tested.

19

Collaborations• Project is conducted in close cooperation with U.S. Industry through the

Advanced Engine Combustion (AEC) / HCCI Working Group.– Ten OEMs, Five energy companies, Four national labs, & Several universities.

• LLNL: Support chemical-kinetic mechanism development, Pitz et al.

• SNL: 1) Collaborate on ethanol HCCI with SI alt.-fuels lab, Sjöberg et al.2) Collaborative project on LES modeling of HCCI, Oefelein et al.

• General Motors: Bi-monthly internet meetings ⇒ in-depth discussions.– Support GM modeling of boosted HCCI and TS with data and discussions.

• U. of Michigan: Support modeling TS, boundary-layer devel. & heat transfer.

• U. of New South Wales: Support modeling of ethanol-fueled HCCI.

• Chevron: Funds-In project on advanced petroleum-based fuels for HCCI.

• JBEI (Joint BioEnergy Institute): Funds-In project on 2nd generation biofuel, iso-pentanol, for HCCI.

• SNL-LDRD: Funds-In project on biofuels produced by fungi ⇒ collab. with researchers in basic chemistry (C. Taatjes et al.) & Biofuels (M. Hadi et al.).

20

Future WorkThermal Stratification• Complete current parametric study using side-view imaging to determine the

of the effects of engine speed and intake temperature on TS.

• Extend parametric study to include: 1) independent variation of firedeck and piston-top temps, & 2) enhancing TS through increased turbulent convection.– Continue collaborations with J. Oefelein et al. on use of LES modeling of TS.

High-Efficiency, Boosted HCCI• Expand studies of PFS for boosted gasoline-fueled HCCI: 1) effects of

engine speed and load, & 2) improved mixture formation to improve stability.– Optical imaging of fuel distribution to assist improved mixture formation.

• Explore additional methods for increasing thermal efficiency of boosted HCCI ⇒ Fuel effects, compression ratio, and Miller cycle.

Support of HCCI Modeling• Continue collaborations with GM-research & U of Mich. on HCCI modeling.• Continue to collaborate with LLNL on improving chemical-kinetic

mechanisms of single components and gasoline-surrogate mixture.

21

Summary• Improvements to side-view imaging provide T-maps showing both bulk-gas

and boundary-layer thermal stratification (TS) simultaneously.

• No evidence of consistent flows transporting colder gas from near-wall regions into central bulk gas.

• Bulk-gas TS (which controls max. PRR) appears to result from in-cylinder turbulence producing turbulent structures extending from the wall.– Most cold pockets in bulk gas are structures attached to firedeck or piston top.– Bulk-gas TS reaches a maximum at TDC.

• Reducing Tin increases therm-eff. by reducing required EGR & heat transfer.

• Gasoline autoignition becomes strongly φ-sensitive with boost. ⇒ Enables large reduction in HRR and PRR with partial fuel stratification (PFS).

• PFS significantly increases high-load limit of gasoline-fueled, boosted HCCI.

• PFS also effective for increasing thermal efficiencies of boosted HCCI.– NOx and soot emissions were ultra-low for all PFS conditions studied.

• Ethanol shows “inverse” φ-sensitivity due to strong thermal effects. ⇒ Allows PFS to reduce HRR, but benefit is much less than for boosted gasoline.

HCCI and Stratified-Charge CI Engine Combustion … and Stratified-Charge CI Engine Combustion Research May 10, 2011 – 9:30 a.m. U.S. DOE, Office of Vehicle Technologies Annual Merit

Documents