A University Consortium on Low Temperature Combustion (LTC) … · 2014. 3. 20. · MIT Camless Engine (Boost and Mode Transitions) UCB Multi-cylinder Engine (Boost and Controls)

LTC University Consortium

A University Consortium on Low Temperature Combustion (LTC) for High Efficiency,

Ultra-Low Emission Engines

AcknowledgementsDOE LTC Consortium project DE-FC26-06NT42629Sandia National LabsLawrence Livermore National LaboratoriesIndustry Partners: Borg Warner, Bosch, BP, Ford Motor Company, General Motors

ParticipantsUniversity of Michigan (UM): Dennis Assanis (PI), Aris Babajimopoulos, Zoran Filipi, Hong Im, George Lavoie, Margaret WooldridgeMassachusetts Institute of Technology (MIT): Wai ChengStanford University (SU): Chris Edwards, Chris GerdesUniversity of California, Berkeley (UCB): Jyh-Yuan Chen, Robert Dibble

This presentation does not contain any proprietary or confidential information.

ace_19_assanis

DOE Vehicle Technologies Program Annual Merit ReviewWashington, D.C., May 18-22, 2009


OverviewTIMELINE• Start – Jan 2006• Finish – June 2009• 85% finished

BUDGET• Total project funding ($3,748k)• Funding FY08 ($1,240k)• Funding FY09 ($947k)

BARRIERS ADDRESSED• Extend operating range of LTC• Develop strategies for exploiting LTC

for optimal FE benefits• Investigate ignition timing control• Explore fuel effects on LTC operation

UNIVERSITY PARTNERS• UM (lead)• MIT• UCB• STANFORD

COLLABORATION• Sandia National Lab• Lawrence Livermore

National Laboratories• General Motors• Borg Warner• Bosch• Ford Motor Company• BP

2


Objectives

• Expand the operating range of LTC engines at both high and low loads

• Investigate methods of achieving improved combustion control

• Investigate effects of stratification on autoignition and combustion

• Determine kinetics of alternative fuels and blends and optimize HCCI operation with such fuels.

0

2

4

6

8

10

12

14

0 1000 2000 3000 4000 5000

NM

EP (b

ar)

RPM

SI MAX LOAD

FTP RANGE

BASE

RANGE

TURBOCHARGING

DILTC

3


Milestones

4

ENGINE MODELINGKiva – burn correlationsDevelop GT modelApplications

HCCI LOAD LIMITSEstablish N.A. limitsLow load – DI effects

-- Kinetic modelingHigh Load - Boosted

WALL EFFECTSTwall/Tin effectsDeposit effects

SPARK ASSISTAcquire images with SAModel lean, high T lam. FlamesDevelop KIVA SACI model

DI - STRATIFIED CHG MODELSFlamelet modelingDNS studies

CHEMISTRY/KINETICSMeasure Ignition delay for biofuelsValidate LLNL model for small estersSampling valve experiments

YEAR 1 YEAR 3YEAR 2


Approach

Remapping Algorithm

11

22

33

44

Temperature zonedivided into ϕ zones

Shaded part:3rd Temperature zone (15-35%)

ϕHigh

Low

TempHigh

Low

Temperature and ϕ distributions

Head

PistonHead

Piston

User subroutinesCombustionHeat transferNOx

Twall

FuelInjector

(DI)

Valve Actuation Strategy

Exh. Gas

Hot Air

Cyl.Block

FastFlap

Valves

Cold Air

KIVA-MZ

Virtual Engine in GT-Power

Optical Engine

Multi-Cylinder ControlDI/VVA Strategies

Boosted HCCISA-HCCI

0

1

2

3

4

5

6

0 2 4 6 8

IMEP

(bar

)

CA50 (deg ATC)

1.7 bar

1.0 bar

GASOLINE

ETHANOL

Fuel Ignition Properties

Fuel Ignition Properties

5

HeatFlux

FFVAEngine


Unique Range of Engine FacilitiesUM Optical Engine

(SA-HCCI and Fuels)UM Heat Transfer Engine(Thermal Management)

MIT Camless Engine(Boost and Mode Transitions)

UCB Multi-cylinder Engine(Boost and Controls)

UM Camless Engine(Multi-Mode Combustion)

6

Stanford Camless Engine(DI and Controls)


-360 -240 -120 0 120 240 360Crank Angle [deg]

PCYL

1 [b

ar]

0

5

10

15

20

25

30

35

40

45

50

55

60

65

[bar

]PM

AN1

PEXH

1

0.0

0.5

1.0

1.5

2.0

PEXH1

PMAN1

[mm

]Va

lve3

_Va

lve2

_Va

lve1

_Va

lve0

_

0

2

4

6

8

10

CoV%MaxMeanMin

IMEP1bar

3.874.354.123.53

AI50_1deg

13.8621.0015.3211.50

PMAN1bar

8.521.231.01

0.865

PEXH1bar

3.881.491.261.13

PMAX1bar

10.0238.7032.5424.63

PCYL1bar

94.2333.3610.080.622

RMAX1bar/deg

36.684.802.521.28

APMAX1deg

16.2920.9017.670.500

0

300

600

900

1200

1500

1800

2100

2400

2700

3000

3300

3600

3900

4200

SPEE

D [r

pm]

SPEEDrpm

1804.5

Integrated high fidelity model and camlessengine test cell for HCCI FE assessment

7

0

2

4

6

8

10

12

14

0 1000 2000 3000 4000 5000N

MEP

(bar

)RPM

A (+8% FE)

B (+11% FE)

MAX (+20% FE)

HCCI

WOT

FTP

0

2

4

6

8

10

0 100 200 300 400 500 600 700

IntakeExhaust

Lift

(mm

)

CAD (deg)

Rebreathing

CAD (deg)

Lift

(mm

)

INTEXH

Recompression

Lift

(mm

)

CAD (deg)

EXHINT

VALVESTRATEGY TYPE GAIN

A Rebreathing ~8 %B Recompression ~11 %

MAX Max. potential ~20 %

MODEL PROJECTIONS

• Developed a fundamentally based HCCI combustion simulation model for use with GT-Power®.

• Model was used to compare different valve strategies for a naturally aspirated HCCI engine subject to NOx, knock and misfire constraints.

• Camless SCTE has been setup and is being used for model validation and strategy assessment.

User subroutinesCombustionHeat transferNOx

Twall

FuelInjector

(DI)

Valve Actuation Strategy

A B

Fully flexible valveactuation (mechanism bySturman Industries)


Explored effect of intake pressure on high load limit

• Used GT Power model with UM developed combustion correlation to investigate the potential of increased intake pressure for extending high load range.

• Ringing intensity criterion can be satisfied by decreasing Φ as boost pressure is increased

Pin0

10

20

30

40

50

60

70

80

90

120

110

100

‐90

‐80

‐70

‐60

‐50

‐40

‐30

‐20

‐10 0 10 20 30 40 50 60 70 80 90 100

110

120

Pressure (bar)

y

1.2 bar

1.4 bar

1.6 bar

1.8 bar

2.0 bar

SuperchargerWork

Crank Angle (deg)

IntakeManifold

TubeBundle

4-to-1ExhaustManifold

Supercharger+ Bypass+ Throttle

Intercooler+ Controller

‐360 ‐300 ‐240 ‐180 ‐120 ‐60 0 60 120 180 240 300 360

Lift

CAD

Scaled Lift and DurationScaled Intake

Scaled Exhaust

Exhaust SI lift

Intake SI LiftRecompressionfor residual gas retention and combustion phasing control

0.30

0.35

0.40

0.45

0.50

0.55

0.60

0.65

0.80

1.00

1.20

1.40

1.60

1.80

2.00

2.20

3 4 5 6 7 8 9

Equ

ival

ence

Rat

io (-

) Intake Pressure (bar)

IMEP (bar)

Φ

Φ' = Φ(1-RGF)

Pin

2

maxmax

max

1Ringing intensity2

dPdt RTP

βγ

γ

⎛ ⎞⎜ ⎟⎝ ⎠≈

8


Demonstrated increased load limitin engine experiments

• Net IMEP’s of above 6 bar achieved in the lab• CA50 must be retarded to decrease ringing

intensity, while avoiding misfire• Further load gains may be possible through

leaner operation at higher boost 2

maxmax

max

1Ringing intensity2

dPdt RTP

βγ

γ

⎛ ⎞⎜ ⎟⎝ ⎠≈

0

2

4

6

8

10

12

14

0 1000 2000 3000 4000 5000

NM

EP (b

ar)

RPM

SI MAX LOAD

FTP RANGE

BASE

RANGE

TURBOCHARGING

DILTC

9

VW 1.9L TDI engine (CR=17:1), 1800 RPM

Ringing Indexlimit

Higher loadspossible here

More fuel, but even more air!


Extended low load limit with DI during NVO• Low load extended to 1 bar NMEP by

injecting fuel during negative valve overlap (NVO) and leaner operation to induce recompression reactions

• Observed advanced combustion phasing and better cycle stability

• Model studies show maximum effect on ignition occurs with Φ ~ 0.8 and moderate exothermicity

0

2

4

6

8

10

12

14

0 1000 2000 3000 4000 5000

NM

EP (b

ar)

RPM

SI MAX LOAD

FTP RANGE

BASE

RANGE

TURBOCHARGING

DILTC

Negative Valve Overlap(NVO)

EXPERIMENT

MODELED REACTIONS DURING NVO

10

Extended low limit

RECOMPRESSIONANGLE OF PEAK PRESSURE

NVO

TDC

Φ ~ 0.8

Φ ~ 1.0

Φ ~ 1.0

Φ ~ 0.8

NVO INJECTION

EN

ER

GY

FRA

CTI

ON

EN

D O

F N

VO

REFORMATION

EXOTHERMICITY

OPTIMAL Φ


2.0

2.5

3.0

3.5

4.0

4.5

1200 1600 2000 2400

y

IMEP

[bar

]

Engine Speed [rpm]

Combustion chamber covered with deposits, equilibrium thickness

Clean combustion chamber walls

Determined the effect of near-wall thermal conditions on HCCI limits

• Increased combustion chamber wall temperature or presence of the thermal barrier on the wall extends the low load limit

• Surface temperature measurements with fast thermocouples enable determining:– Time-varying boundary conditions, with or without

deposits

• Coupled heat transfer experiments with CFD modeling of boundary layer effects to provide: – In-depth insight into thermal stratification– Guidance for developing HCCI range expansion

HCCI Operating RangeSingle-cylinder gasoline HCCI with re-induction of residual, fixed exhaust cam lift, 2000 rpm. Experimental work performed with UM/GM CRL funding

Wall affected zone highly influenced by wall temperature, although thin in can contain large fraction of mass

CO

Temp.

0.0

0.5

1.0

1.5

2.0

2.5

-120 -90 -60 -30 0 30 60 90 120

p1p3p4p5p6

p7p8h1h2Local Average

Hea

t Flu

x [M

W/m

2 ] KIVAEXP

11

GM/UM Collaborative Research Laboratory


Insight gained on spark assisted HCCI (SACI)with optical engine

• Spark assist with high load (Φ ~0.6) and low T is dominated by flame type behavior

• Without spark, images show similar reaction fronts but much slower overall combustion

• Spark assist with low load (Φ ~ 0.4) and high T shows mixed mode combustion

• Without spark, combustion is slower and less stable

12

HIGH LOAD (Φ=0.62); LOW Tint (271C)

LOW LOAD (Φ=0.40); HIGH Tint (321C)

SPK-70 ATC

NO SPK

SPK-5 ATC

NO SPK


Extended model based laminar flame data for application to HCCI and SACI ranges

• Transient flame code (HCT) showed that autoignition and flame propagation are largely independent processes

• Used HCT to establish database of flame speeds beyond range of available experimental data

• Flame speeds ( SL~20-40 cm/s) appear robust enough for flame propagation in SACI and HCCI regions (consistent with optical engine observations)

• KIVA model nearing completion based on wrinkled laminar flame combined with autoignition

13

500

1000

1500

2000

2500

3000

500 1000 1500

TB (K

)TU (K)

1.00.8

0.6

0.4

0.2

0.0

Φ NOx

FLAMELIMIT

BULKQUENCH

EARLYCOMB

KNOCKLIMIT

AUTOIGN

SISACI

HCCI

Metghalchi and Keck, (1982)

Bradleyet al., (1998)

UM HCT Generated

Dataset

0

500

1000

1500

2000

2500

0.000 0.001 0.002 0.003 0.004 0.005

Tempe

rature (K

)

Mass (g/cm^2)

Temperature vs. Mass

steadyflame

mixed modeautoignition

HCT calculation

TEM

P (K

)BURNED

GAS

UNBURNED

FLAME

KIVA SACIMODEL


Ignition delay: case C > case A > case B

Peak heat release: case C > case A > case B

Case A

Case B

Case C

Effects of SOI on mixture formation at TDC:

• Premixed – constant Ф (case A)

• Early SOI – uncorrelated T-Ф fields (Case B)

• Late SOI – negatively-correlated T-Ф fields (Case C)

Initial Conditions

T

Investigated autoignition in LTC engine environments with DNS

• Small scale effects of T-Φ correlations on autoignition and front propagation studied using DNS

• Most stratified case displays most spatially distributed reactions and burns faster, while homogeneous and uncorrelated cases exhibit reaction fronts and longer burn durations

• Domain: 4.1mm x 4.1mm(960 x 960)

• Detailed chemistry of H2• P = 41 atm• Prescribed random turbulence

field (u’ = 0.5m/s, L11 = 0.34mm)• Prescribed random temperature

field (Tmean=1070K, T’=15K)

Normalized HRR

14


iiii

ii

ZY

ZY

tY ωρχρρρ ++

∂∂

=∂∂

+∂∂

&&2

2

2

Radial Distance [cm]

CO

[g](

forS

OI2

40an

d27

0)

CO

[g](

forS

OI3

00)

0 1 2 3 4 5-0.004

-0.002

0

0.002

0.004

0.006

0.008

0

2E-06

4E-06

6E-06

8E-06

1E-05SOI 240SOI 270SOI 300

High stratification: Near WallLow stratification: bulk gas

More S

tratification

CO Contour at TDCWith different stratification level

Studied effect of DI stratification on LTC

• Developed KIVA- RIF-ER model, which considers the effect of evaporation in the reaction space for accurate modeling of LTC/DI combustion

• Demonstrated model capability by matching experiments (Dec,2003)

• Studied effect of stratification on spatial CO production under low load conditions

Effect of evaporation includedin the reaction space

Effect of evaporation

15


Determined ignition properties of biofuels and fuel blends

• Measured ignition delays for 3 small biofuel esters (basic components of typical biofuels) and found a factor of 3 variation in delay times

• Speciation studies of the ester methyl butatnoate (MB) were conducted and the key reaction pathways identified.

• There was excellent agreement between the reaction mechanism developed by Westbrook, Pitz and co-workers at LLNL

• Identified non-linear behavior for methyl trans-3-hexenoate (M3H) and n-heptane blends

• Results indicate that biofuel blends could be designed with targeted levels of HCCI reactivity

M3H

MC

MB

Temperature [K]

0.90 0.95 1.00 1.05 1.101000/T [1/K]

1

100

10

τig

n(m

s)

16

Rapid Compression Facility

2000

0 0.2 0.4 0.6 0.8 1.0Normalized Time [TSAMPLE/τIGN]

Mol

e Fr

actio

n (p

pm)

0

ethene2000


Mol

e Fr

actio

n (p

pm)

0

2000


Mol

e Fr

actio

n (p

pm)

0

etheneLLNL MODELPREDICTIONSOF ETHANE

EXPT’LRCF

CONC.

METHYL BUTANOATE AUTOIGNITION

P

1

10

100

0.0 0.2 0.4 0.6 0.8 1.0

τign (

ms)

Blending Fraction of n-Heptane

NON-LINEARREGION

n-Heptane in M3H


Future Work• Focus on High Pressure – Lean Burn path toward

overall fuel economy gain of 20-40%– Application to downsized and boosted engines– Lean burn and high temperature for thermodynamic

gains• Determine ways to use fuel and thermal

stratification interactions for improved combustion• Explore fuel blends and their effect on combustion

limits• Investigate and demonstrate the benefits of multi-

mode combustion methods (SACI, DI) for combustion control

• Use full range of models developed in previous years (CFD, system, etc) to maintain close connection between experimental work and final, in-vehicle results, including thermal and other transients

~ 30 %FEff Gain

Incr.CR

RICHLEAN

A

B

17


Summary• We have a well developed and balanced approach to the research task

– Full range of modeling tools from fundamental (flame, kinetic, CFD) to system level (GT power)

– Excellent selection of experimental engine facilities (single cylinder, multi-cylinder, rapid compression facility)

– Tools now available to fully focus work on achieving large fuel economy benefits• We have accomplished our objectives for the project so far

– Demonstrated FE potential of candidate valve strategies– Extended low load limit to 1 bar NMEP by DI during NVO– Extended high load limit to 6 bar NMEP by boosted, dilute operation– Demonstrated controllability improvement with spark assist and showed multi-mode

combustion by propagating reaction fronts and autoignition– Obtained fundamental ignition data for biofuels and fuel blends for optimizing fuel/engine

interactions– Used DNS to show that stratification can effect ignition and heat release depending on

nature of T, Φ correlation– Demonstrated combustion control by fast thermal management on multi-cylinder engine

18

Slide Number 1OverviewObjectivesMilestonesApproachUnique Range of Engine FacilitiesIntegrated high fidelity model and camless�engine test cell for HCCI FE assessmentExplored effect of intake pressure on high load limitDemonstrated increased load limit�in engine experimentsExtended low load limit with DI during NVO�Determined the effect of near-wall thermal conditions on HCCI limits��Insight gained on spark assisted HCCI (SACI)�with optical engine��Extended model based laminar flame data �for application to HCCI and SACI ranges �Slide Number 14Slide Number 15Determined ignition properties of biofuels and fuel blendsFuture WorkSummaryMaterial for ReviewersMaterial for ReviewersFY 08 PublicationsSlide Number 22

Documents

A University Consortium on Low Temperature Combustion (LTC) … · 2014. 3. 20. · MIT Camless Engine (Boost and Mode Transitions) UCB Multi-cylinder Engine (Boost and Controls)