Low-Temperature Gasoline Combustion (LTGC) Engine Research
– Previously known as HCCI / SCCI –
John E. DecJeremie Dernotte and Chunsheng Ji
Sandia National Laboratories
June 8, 2016 – 9:00 a.m.U.S. DOE, Office of Vehicle TechnologiesAnnual Merit Review and Peer Evaluation
Program Managers: Gurpreet Singh & Leo Breton Project ID: ACE004
This presentation does not contain any proprietary, confidential, or otherwise restricted information.
Timeline● Project provides fundamental
research to support DOE/Industry advanced engine projects.
● Project directions and continuation are evaluated annually.
Budget● Project funded by DOE/VT:
FY15 – $680kFY16 – $600k
Barriers● Rapid control of LTGC / HCCI
combustion timing● Spark-Assisted LTGC / HCCI● Improved stability / robustness of
LTGC combustion● Advanced fuel-injection strategies● Improved understanding of LTGC
fundamentals
Partners / Collaborators● Project Lead: Sandia ⇒ John E. Dec● Part of Advanced Engine Combustion
working group – 15 industrial partners● General Motors – in-depth collaboration● Cummins – spark-plug cylinder heads● LLNL – support kinetic modeling● Co-Optima Fuels project● Chevron – advanced fuels for LTGC● Sandia LDRD – fuel injection
Overview
Objectives - Relevance
FY16 Objectives ⇒ address barriers, particularly Controls and Robustness
● Performance mapping with new low-swirl, spark-plug capable cylinder head: Compare thermal efficiency (TE) & load range with data from old head.
● Evaluate performance with RD5-87 (typical regular 87 AKI, E10 gasoline) compared to Tier-2 certification gasoline (CF-E0) for premixed (PM) fueling and with partial fuel stratification (PFS).
● CA50 control and improved robustness using Double-DI PFS (DDI-PFS) ⇒ Determine the potential for CA50 control and improved EGR tolerance.
● Initial studies of Spark Assist (SA): Determine CA50 control and intake-temperature (Tin) tolerance at selected conditions.
● Support Modeling: Chemical-kinetics at LLNL and related RCM experiments at ANL, and CFD modeling at GM.
Project objective: to provide the fundamental understanding (science-base) required to overcome the technical barriers to the development of practical LTGC / HCCI engines by industry.
Response to Reviewer Comments● Reviewers made many positive comments. ⇒ We thank the reviewers
● Several comments indicated ⇒ focus less on high efficiency and high loads and more on ways control combustion timing and operation at lower boost.► We have accelerated plans to shift research in these directions, as reflected in the
FY16 objectives (prev. slide) and explained in greater detail below.
Specific comments1. Accelerate installation of spark-plug head and studies of spark assist (SA)
► Several mechanical/technical problems were encountered that delayed installation of head.► Head was installed latter part of FY15, debugged. Initial studies of SA have been conducted.
2. Studies of DDI-PFS should include CA50 control methodologies. ► DDI-PFS has strong potential for rapid CA50 control and for increased robustness.
⇒ We have shifted the focus of DDI-PFS studies to these objectives.
3. Concerns that high boost can be difficult with LTGC ► PFS requires that fuel autoignition be φ-sensitive ⇒ typically greater at higher boost.► Investigated φ-sensitivity over a wide range of boost for CF-E0 and RD5-87
⇒ Found good potential that PFS can provide benefits down to Pin=1.3 bar, better for RD5-87► New studies have been conducted at lower boost ⇒ additional low-boost studies planned.
4. Need to show Combustion Noise Levels (CNL) as well as Ringing Intensity (RI)► CNL values are presented and discussed.
Approach
● Metal Engine► Modify new cylinder heads to install spark-plug (SP) ports.
> Work with Cummins on design, SP port installation, & new pressure transducer (PT) port> In-house modifications to SP-head for Bosch HDEV 5.1 GDI injector (300 bar capable).
► Well-controlled experiments to 1) evaluate SP-head performance, and investigate: 2) DDI-PFS: develop methods of varying fuel stratification to obtain injection-timing
control of CA50, increased CA50 tolerance, and improved stability. 3) Spark-Assist: systematically adjust spark time for CA50 ctrl. & Tin compensation.
● Optical Engine – adaptation of SP-head and installation will follow.● Fuels – Worked with GM to specify a research-grade E10 regular gasoline, RD5-87,
and compare performance with CF-E0. (Prior to recent E10 Tier 3 cert. gas.)● Analytical Techniques – Apply our recently developed techniques to understand:
1) changes in energy-loss distribution, and 2) noise levels, CNL● Computational Modeling: 1) Collaborate with LLNL on kinetic mechanism for
RD5-87, and 2) with GM on CFD modeling for improved understanding of PFS.● Combining techniques provides a better understanding and more-optimal solutions.● Transfer results to industry: 1) physical understanding, 2) improved models
Overall Approach: Use a combination of metal- and optical-engine experiments, analysis & modeling to build a comprehensive understanding of LTGC processes.
Approach – Milestones● September 2015
Complete installation and initial testing of new low-swirl cylinder head with spark-assist capability.
● December 2015Map performance of SP-head (Head #2) over a range of operating conditions and compare with previous head (Head #1).
● March 2015Complete installation of spark ignition system and initial study of spark-assisted (SA) LTGC.
● June 2016Present an overview of project accomplishments and directions at the DOE Annual Merit Review.
● September 2016Map the operating range for effective DI-PFS with E10 regular gasoline at a compression ratio of 14:1 (plan to switch soon from current CR = 16:1 to 14:1).
6
Unless noted: Ringing ≤ 5 MW/m2 & spd = 1200 rpmNOx & soot emiss. more than 10x below US-2010
Sandia LTGC Engine Laboratory
All-Metal Engine
Optical Engine
Optics Table
Dynamometer
Intake Plenum
Exhaust Plenum
Water & Oil Pumps & Heaters
Flame Arrestor
● Matching all-metal & optical LTGC 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 = 16:1, switch to 14:1
Overview of Accomplishments● Completed installation and shakedown testing of new spark-plug capable,
low-swirl cylinder head (Head #2).
● Conducted performance mapping of Head #2 and comparisons with Head #1 for both premixed & Early-DI fueling ⇒ TE, high-load limits, CNL, etc.– Applied energy-loss analysis tools (developed in FY15) to understand differences.
● Evaluated performance of a research-grade regular 87-AKI, E10 gasoline (RD5-87) and compared to high-octane, E0 certification gasoline (CF-E0).
● Demonstrated CA50 control over a wide range by varying injection timingfor a DDI-PFS fueling method: – Retard late-DI timing incr. strat. adv. CA50
● Showed that DDI-PFS can also substantially increase robustness (EGR & CA50 tolerance) and increase stability for an extended load range.
● Demonstrated Spark-Assisted (SA) LTGC for CA50 control and increased tolerance to variation in Tin (compensate for Tin variation).
● Collaborated with LLNL on development of a kinetic mechanism for RD5-87 and related RCM measurements at ANL, and with GM on CFD modeling.
Low-Swirl Spark-Plug Head ⇒ “Head #2”● Worked with Cummins to design SP
capability and fabricate.– SP port in location of original
D = 10 mm PT (AVL QC34C)
● Install new PT port through fire-deck– Very small, D = 5 mm (AVL GH15D)– For CI studies, 2nd GH15D in SP port.
● Problems w/ small PT, not all are durable.
● Both heads are low-swirl, but:– Head #1, custom anti-swirl plate directs
helical port flow against tangential port to create a counter-swirling flow.
– Head #2, ports designed to give low swirl⇒ thought to produce tumble flow.
● Central-mount Bosch HDEV 5.1 GDI injector ⇒ 300 bar capable.
● Same valves / camshaft / rocker assembly for both heads.
2.45
4.39.5
3
33.68
Round 0.5
5
Exhaust
Intake
NGKDFE
Iridium
Head #1Counter-swirling
streams
Head #2Assumed tumble
air motion
PT Port
9
45.0
45.5
46.0
46.5
47.0
47.5
48.0
48.5
49.0
49.5
50.0
0.25 0.30 0.35 0.40 0.45
Ther
mal
Effi
cien
cy [%
]
Charge-Mass Equiv. Ratio [φm]
PreMixed, RI = 5 MW/m2
1200 RPMPin = 240 kPa, Tin = 60°C
RI = 5.39
Head #1
Head #2
Thermal Eff. (PM) – Spark-Plug Head #2● Initial testing of Head #2 used:
– CF-E0 ⇒ large database for Head #1– Premixed (PM) fueling to eliminate
differences due to fuel inject & mixing.
● φm ≥ 0.34: TE with Head #2 is just slightly lower (~0.2 %-units).
● φm < 0.34: greater TE loss w/ Head #2● Cause is not well understood:
– Combst. Eff (CE) and CA50 are similar– EGR requirement & γ also similar
● Analysis shows increased HT with Head #2 is the most likely explanation.– Possibly high-tumble flow breaks down
near TDC and increases HT.– Greater at low φm
since CA50 is closer to TDC.
● Is counter-swirl better for low HT?
stoichm AF
CF)/(
)/(=φ
PreMixed, CF-E0, Pin = 2.4 bar, RI = 5 MW/m2
362
364
366
368
370
372
374
376
378
95.5
96.0
96.5
97.0
97.5
98.0
98.5
99.0
99.5
0.25 0.30 0.35 0.40 0.45
CA5
0 [°
CA]
Com
bust
ion
Effic
ienc
y [%
]
Charge-Mass Equiv. Ratio [φm]
Head #1, CEHead #2, CEHead #1, CA50Head #2, CA50
Head #1 Head #2
Oxygen limited
10
94.094.595.095.596.096.597.097.598.098.599.099.5
100.0
0.25 0.30 0.35 0.40 0.45
Com
bust
ion
Effic
ienc
y [%
]
Charge-Mass Equiv. Ratio [φm]
Head #1 - 70° injector
Head #2 - 90° injector
PM, Head #1
PM, Head #2
Double injection:70% DI @ 60°CA+ 30% DI @ 150°CA
45.0
46.0
47.0
48.0
49.0
50.0
51.0
0.25 0.30 0.35 0.40 0.45
Ther
mal
Effi
cien
cy [%
]
Charge-Mass Equiv. Ratio [φm]
E-DI @ 60°CA, RI = 5 MW/m2
1200 RPMPin = 240 kPa, Tin = 40°C
Head #2
Head #1
Thermal Eff. (Early-DI) – Spark-Plug Head #2
● Compare heads, Early-DI @ 60° CA– Tin = 40°C vs. 60°C for Premixed– Injection Press = 120 bar, both heads
● Overall TE higher than PM mainly due to lower Tcharge ⇒ higher γ & lower HT.
● φm ≤ 0.4: Trends similar to PM, but– For φm ≤ 0.35: TE reduction with Head #2
slightly larger for E-DI due to lower CE with Head #2 ⇒ higher CO.
● φm > 0.4: TE of Head #2 falls below Head #1, rapid drop in CE ⇒ higher CO
● Increased CO at low and high φm indicate a less well-mixed charge with Head #2.– Low φm overly lean zones make CO– High φm rich zones make CO – high EGR
● Counter-swirl improves mixing for Early-DI fueling with Head #1.
stoichm AF
CF)/(
)/(=φ
Early-DI, CF-E0, Pin = 2.4 bar, RI = 5 MW/m2
94.094.595.095.596.096.597.097.598.098.599.099.5
100.0
0.25 0.30 0.35 0.40 0.45
Com
bust
ion
Effic
ienc
y [%
]
Charge-Mass Equiv. Ratio [φm]
Head #1 - 70° injector Head #2 - 90° injector Head #2 - 90° injector - S-DI (repeat) Head #2 - 90° injector - D-DI PM, Head #1 PM, Head #2
Double injection:70% DI @ 60°CA+ 30% DI @ 150°CA
11
● Max. load for PM fueling with CF-E0 nearly identical for two heads, all Pin.– Oxygen limited Pin > 2.4 bar (CF-E0).
● RD5-87, PM is similar to CF-E0.– Max. load slightly greater, Pin ≤ 2.0 bar.⇒ Higher φm for RI = 5 MW/m2.
– Max. load is less at Pin=2.4 bar⇒ More reactive, requires more EGR,
becomes O2-limited at Pin < 2.4 bar.
● Early-DI fueling: Max. load quite similar for two heads, all Pin.– Highest load at Pin = 3.0 bar⇒ 17.2 bar IMEPg for Head #2 vs.⇒ 17.7 bar IMEPg for Head #1
● Combustion Noise (CNL) is similar for all max. load curves, for RI = 5MW/m2
– Close to high end of diesel CNL range.– Could reduce CNL by small CA50 retard
5 dBA reduction for 0.8 %-units less TE
High Load Limit as a Function of Boost
6065707580859095
100105110
2 4 6 8 10 12 14 16 18 20
CNL
[dBA
]
IMEPg [bar]
PM, CF-E0, Head #1PM, CF-E0, Head #2DI-60, CF-E0, Head #1DI-60, CF-E0Head #2
Typical range for modern diesel engines(E. Kurtz, Ford, AEC 2015)
Pin = 1.0 bar
Pin = 3.0 bar, PMPin = 3.0 bar DI-60
RI = 5 MW/m2 (except as noted in upper plot)
2
4
6
8
10
12
14
16
18
20
0.8 1.2 1.6 2.0 2.4 2.8 3.2
Max
imum
IMEP
g[b
ar]
Intake Pressure [bar]
PM, CF-E0, Head #1
PM, CF-E0, Head #2
Knock/Stability limited O2 limited
1200 RPMRI = 5 MW/m2 (except as noted)
RI = 3
2
4
6
8
10
12
14
16
18
20
0.8 1.2 1.6 2.0 2.4 2.8 3.2
Max
imum
IMEP
g[b
ar]
Intake Pressure [bar]
PM, CF-E0, Head #1PM, CF-E0, Head #2PM, RD5-87, Head #2
Knock/Stability limited O2 limited
1200 RPMRI = 5 MW/m2 (except as noted)
RI = 3
RD5-87 O2 limited
RI = 3.5
2
4
6
8
10
12
14
16
18
20
0.8 1.2 1.6 2.0 2.4 2.8 3.2
Max
imum
IMEP
g[b
ar]
Intake Pressure [bar]
PM, CF-E0, Head #1PM, CF-E0, Head #2PM, RD5-87, Head #2DI-60, CF-E0, Head #1DI-60, CF-E0, Head #2
Knock/Stability limited O2 limited
DI-60, Head #1 - RI = 31200 RPMRI = 5 MW/m2 (except as noted)
DI-60, Head #2 - RI = 3
12
Injection-Timing/PFS to Control LTGC
-0.02
0.03
0.08
0.13
0.18
0.23
0.28
350 355 360 365 370 375 380 385 390
HRR/
THR
[1/o C
A]
Crank Angle [°CA]
Premixed, Tin=60CSingle-DI, Tin=73CSingle-DI, Tin=40CD-DI, 8%@305D-DI, 14%@305D-DI, 20%@305D-DI, 25%@305D-DI, 30%@305D-DI, 35%@305
Tin = 40°C, Pin = 2.4 bar, φm = 0.4, 1200 rpm
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 bar
Pin = 2.4 bar
Continuous lines: PMDashed-lines: Std-PFS
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 barPin = 2.4 barPin = 2.0 barPin = 1.8 barPin = 1.6 barPin = 1.3 barPin = 1.0 bar
Continuous lines: PMDashed-lines: Std-PFS
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 barPin = 2.4 barPin = 2.0 barPin = 1.8 barPin = 1.6 barPin = 1.3 barPin = 1.0 bar
Continuous lines: PMDashed-lines: Std-PFS
CF-E0RI = 5 MW/m2
EGR adjustedto hold CA50 constant
● If the fuel’s autoignition timing varies with the local in-cyl. φm, said to be φ-sensitive⇒ richer regions autoignite faster.
● Partial fuel stratification (PFS) can be used to provide several benefits.– Reduced HRR for higher loads & higher TE.
⇒ Shown in previous years.– Combustion-timing control– Increased robustness, i.e. tolerance to
variation in EGR and CA50
● Std-PFS = most Premixed + late DIDouble-DI PFS = most Early-DI + late DI⇒ late-DI timing & fraction adjusts strat.
● For what Pin range are fuels φ-sensitive? ⇒ Direct measurement very tedious.
● Use CA50 adv. for RI = 5 MW/m2 with std-PFS vs. PM as a measure of φ-sensitivity.– Here std-PFS = 90% PM + 10% at 310° CA.
Injection-Timing/PFS to Control LTGC
-0.02
0.03
0.08
0.13
0.18
0.23
0.28
350 355 360 365 370 375 380 385 390
HRR/
THR
[1/o C
A]
Crank Angle [°CA]
Premixed, Tin=60CSingle-DI, Tin=73CSingle-DI, Tin=40CD-DI, 8%@305D-DI, 14%@305D-DI, 20%@305D-DI, 25%@305D-DI, 30%@305D-DI, 35%@305
Tin = 40°C, Pin = 2.4 bar, φm = 0.4, 1200 rpm
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 bar
Pin = 2.4 bar
Continuous lines: PMDashed-lines: Std-PFS
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 barPin = 2.4 barPin = 2.0 barPin = 1.8 barPin = 1.6 barPin = 1.3 barPin = 1.0 bar
Continuous lines: PMDashed-lines: Std-PFS
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 barPin = 2.4 barPin = 2.0 barPin = 1.8 barPin = 1.6 barPin = 1.3 barPin = 1.0 bar
Continuous lines: PMDashed-lines: Std-PFS
CF-E0
360
362
364
366
368
370
372
374
376
0.25 0.30 0.35 0.40 0.45 0.50
CA
50 [°
CA
]
Charge-Mass Equiv. Ratio [φm]
Pin = 2.4 barPin = 2.0 barPin = 1.6 barPin = 1.3 barPin = 1.0 bar
Continuous lines: PMDashed-lines: Std-PFS
RI = 5 MW/m2
RD5-87
EGR adjustedto hold CA50 constant
● If the fuel’s autoignition timing varies with the local in-cyl. φm, said to be φ-sensitive⇒ richer regions autoignite faster.
● Partial fuel stratification (PFS) can be used to provide several benefits.– Reduced HRR for higher loads & higher TE.
⇒ Shown in previous years.– Combustion-timing control– Increased robustness, i.e. tolerance to
variation in EGR and CA50
● Std-PFS = most Premixed + late DIDouble-DI PFS = most Early-DI + late DI⇒ late-DI timing & fraction adjusts strat.
● For what Pin range are fuels φ-sensitive? ⇒ Direct measurement very tedious.
● Use CA50 adv. for RI = 5 MW/m2 with std-PFS vs. PM as a measure of φ-sensitivity.– Here std-PFS = 90% PM + 10% at 310° CA.
● Both fuels φ-sensitive from Pin = 2.4–1.3 bar⇒ RD5-87 more φ-sensitive, all φm s & Pins.
0
2
4
6
8
10
70
75
80
85
90
95
180 200 220 240 260 280 300 320
Ring
ing
Inte
nsity
[MW
/m2 ]
CNL
[dBA
]
2nd DI timing (30%) [°CA]
Ringing Intensity Limit
CNL upper range for modern diesel engines IMEPg = 12 bar (E. Kurtz, Ford , AEC 2015)
CNL lower range
Combustion Noise (CNL)
CA50 Control with Injection-Timing
0
2
4
6
8
10
12
14
16
18
20
368
369
370
371
372
373
374
375
376
377
378
180 200 220 240 260 280 300 320
Ring
ing
Inte
nsity
[MW
/m2 ]
CA
50 [o C
A]
2nd DI timing [°CA]
D-DI, 30% late-DI
S-DI, CA50 = 373.3 RI = 2.5 MW/m2
COV of IMEPg = 3.0%
COV of IMEPg = 1.9%
RI = 5 MW/m2
0
2
4
6
8
10
12
14
16
18
20
368
369
370
371
372
373
374
375
376
377
378
180 200 220 240 260 280 300 320
Ring
ing
Inte
nsity
[MW
/m2 ]
CA
50 [o C
A]
2nd DI timing [°CA]
D-DI, 30% late-DID-DI, 20% late-DI
S-DI, CA50 = 373.3 RI = 2.5 MW/m2
COV of IMEPg = 3.0%
COV of IMEPg = 1.9%
RI = 5 MW/m2
RD5-87, Pin = 2.0 bar, Tin = 40°C, φm = 0.4● Apply Double-DI (DDI) PFS to control CA50.● Procedure:
1. Set initial conditions ⇒ adjust CA50 to give RI=2.5 MW/m2 for single, Early-DI injection.
2. Switch to DDI with 70% Early-DI at 60°CA &30% late-DI with variable timing (70/30%).
3. Hold EGR and Tin constant while sweeping late-DI timing.
● Late-DIs from 200 – 280° CA retards CA50 compared to Single-DI at 60°CA (S-DI-60).– Indicates better mixing than S-DI, which
already gives some PFS. ⇒ RI < 2.5 MW/m2
● Late-DIs from 280 – 300° CA advance CA50 significantly due to greater stratification.⇒ RI = 2.3 – 6.1 MW/m2
● CA50 was adjusted 6.7° CA with 70/30%(4.5° COV-IMEPg = 1.9% to RI = 5 MW/m2)
● With DDI-80/20%, CA50 ctrl. range 8.6° CA● CNL trend is similar to RI ⇒ below upper range
for diesels for most of the sweep, RI ≤ 3.5 MW/m2.
10.4
10.5
10.6
10.7
10.8
10.9
11.0
11.1
11.2
11.3
364 366 368 370 372 374 376 378
Inta
ke O
2[m
ole
%]
CA50[°CA]
D-DI, 30%S-DI
Double-DI0.28 O2%-units
Single-DI0.08 O2%-units
0
1
2
3
4
5
6
7
8
02468
101214161820
364 366 368 370 372 374 376 378
CO
V of
IMEP
g[%
]
Rin
ging
Inte
nsity
[MW
/m2 ]
CA50[°CA]
D-DI 70/30% at 305° CA S-DI-60° CA
Ringing limit &Stability limit
D-DI, 30%4.3° CA
S-DI1.8° CA
Increasing Robustness with DDI-PFSCF-E0, Pin = 2.4 bar, Tin = 40°C, φm = 0.4● Our general range of acceptable operation
from “knock” to “poor stability” is from RI = 5 MW/m2 to COV-IMEPg = 2%.
● Sweep EGR at constant Tin to shift CA50 across a wide range for S-DI and DDI.⇒ Use intake O2 as a metric for EGR.
● S-DI-60: CA50 tolerance = 1.8° CA⇒ EGR must vary ≤ 0.08 O2%-units.
● DDI, 30% at 305°CA: CA50 tol. = 4.3°CA⇒ EGR can vary ≤ 0.28 O2%-units
● DDI-PFS greatly increases tolerance to non-ideal CA50 and EGR levels.
0
1
2
3
4
5
Allo
wab
leC
A50
Var
iatio
n [°
CA
]
S-D
I
DD
I30%
@ 3
05°
DD
I 14%
@ 3
05°
CA50
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
Allo
w. V
ar. I
ntak
e O
2[%
-uni
ts]
S-D
I
DD
I30%
@ 3
05°
DD
I 14%
Intake O2
Increasing Stability with DDI-PFS, Pin = 1.6 bar● Both Head #1 and Head #2 show reduced
stability for Early-DI (S-DI-60) at Pin=1.6 bar⇒ Cause is not understood.
● Maximum fueling rate (φm) is significantly reduced compared to PM or S-DI-60 at other Pins.– Becomes unstable if φm is increases, and
quickly runs away to knock or misfire.
● With RD5-87, max. φm with S-DI-60 is even lower than with CF-E0.
● Apply DDI-PFS with an relatively early “late-DI” timing ⇒ 80% at 60° + 20% at 200°CA
● DDI-80/20%-200 greatly increases stability, allowing a substantial load increase.⇒ φm increased from 0.34 to 0.42
● Moreover, still stable at φm = 0.42, so further increases are possible.
● Even greater increases may be possible with optimization of DDI fueling strategy.
700
750
800
850
900
950
1000
1050
0.25 0.30 0.35 0.40 0.45 0.50
IMEP
g[k
Pa]
Charge-Mass Equiv. Ratio [φm]
D-DI, 80%@60, 20%@200Single-DI-60
RI = 3 MW/m2
D-DI 80/20%, Still stable,higher loads possible
RI = 3 MW/m2
Stability limitfor S-DI-60
Max. φm at Pin = 1.6 bar, DDI-80/20% vs. S-DI
RD5-87
0.320.340.360.380.400.420.440.460.48
0.8 1.2 1.6 2.0 2.4 2.8 3.2
Max
imum
φm
Intake Pressure [bar]
CF-E0, PM, Head #1 CF-E0, PM, Head #2 CF-E0, DI-60, Head #1 CF-E0, DI-60, Head #2 RD5-87, DI-60, Head #2
0.320.340.360.380.400.420.440.460.48
0.8 1.2 1.6 2.0 2.4 2.8 3.2
Max
imum
φm
Intake Pressure [bar]
CF-E0, PM, Head #1 CF-E0, PM, Head #2 CF-E0, DI-60, Head #1 CF-E0, DI-60, Head #2 RD5-87, DI-60, Head #2RD5-87, DDI 80/20%-200
Max. φm at Max. Load Point, 1200 rpm
0
2
4
6
8
10
12
14
16
18
180
200
220
240
260
280
300
320
340
360
95 100 105 110 115 120 125
CO
V-IM
EPg
[%]
Spar
k Ti
min
g [°
CA
]
Tin [°C]
No spark
0
2
4
6
8
10
12
14
16
18
180
200
220
240
260
280
300
320
340
360
95 100 105 110 115 120 125
CO
V-IM
EPg
[%]
Spar
k Ti
min
g [°
CA
]
Tin [°C]
No spark
Spark Assist
Spark Timing
Spark-Assist for LTGC Control, Pin = 1 bar● Spark-assist (SA) is a promising control
method, Pin=1 bar & lower boost (limit=?)● Complements injection-timing/PFS
control at higher Pin ⇒ fuel is φ-sensitiveRobustness: φ = 0.42, PM fueling● For CI only (no SA), ∆Tin = 3.7°C from
RI = 5 MW/m² to COV-IMEPg = 2%– ∆Tin = 3.7°C gives a ∆CA50 = 7° CA
● For SA + CI, can reduce Tin & maintain CA50 and RI by advancing spark-timing.– Limited by large cycle-to-cycle variations;
COV suddenly becomes >> 2%.> Variability in early-flame propagation
– ∆Tin = 21°C
● Spark assist greatly increases tolerance to Tin variation, from 3.7 to 21°C.
● No significant change in CA50, RI, orCE. Slight decrease in NOx ⇒ lower Tin
Maintain RI = 5 MW/m2 w/ Spark
PM, φm = 0.42
365
366
367
368
369
370
371
372
373
374
375
0
1
2
3
4
5
6
7
8
9
10
95 100 105 110 115 120 125
CA
50 [°
CA
]
Rin
ging
Inte
nsity
[MW
/m²]
Tin [°C]
No spark
No spark
COV 2% limit
365
366
367
368
369
370
371
372
373
374
375
0
1
2
3
4
5
6
7
8
9
10
95 100 105 110 115 120 125
CA
50 [°
CA
]
Rin
ging
Inte
nsity
[MW
/m²]
Tin [°C]
No spark
No spark
COV 2% limit
-10
0
10
20
30
40
280 300 320 340 360 380 400
Heat
Rel
ease
Rat
e [J
/˚CA
]
Crank Angle [˚CA]
Pin = 1.0 barRI ≈ 5 MW/m2
0
50
100
150
200
250
300
350
400
-10
0
10
20
30
40
50
60
70
310 320 330 340 350 360 370 380 390 400
Heat
Rel
ease
Rat
e [J
/˚CA
]
Cyl
inde
r Pre
ssur
e [b
ar]
Crank angle [˚CA]
No spark, Tin = 122 °CSpark tim. = 334.6 CAD, Tin = 115 °CSpark tim. = 329.6 CAD, Tin = 110 °CSpark tim. = 324.5 CAD, Tin = 105 °CSpark tim. = 319.5 CAD, Tin = 100 °C
Pin = 1.0 barRI ≈ 5 MW/m2
Flame Propagation Effect on HRR, φ = 0.42
● First part of HR associated with flame propagation contributes a significant fraction of the total HR.– Up to about 15%
● Compression heating caused by the flame combustion appears to compensate for decrease in Tin– Effect is similar to the ITHR for
boosted operation with CI.
● Can the flame propagation allow CA50 to be retarded further while maintaining robust combustion (COV-IMEPg < 2%)?
● How much control over CA50 does SA provide?
366367368369370371372373374375376377
325 330 335 340 345 350 355
CA
50 [°
CA
]
Spark Timing [°CA]
Tin = 117 deg-CTin = 107 deg-C
RI = 5 MW/m2
COV IMEPg = 2 %
RI = 5 MW/m2
COV IMEPg = 2 %
RI = 5 MW/m2
COV IMEPg = 2 %
RI = 2.8 MW/m2
RI = 0.8 MW/m2
CA50 Control with Spark AssistReminder: Tin = 123°C for no spark, RI = 5 Lowest Tin with spark = 102°C Max. CA50 retard w/o spark = 374° CA
(limited by COV-IMEPg = 2%)
● Spark timing swept at two Tins:– 117°C ⇒ if no spark, COV-IMEPg > 5% – 107°C ⇒ if no spark, no combustion
● Retard CA50 by retarding spark timing,from RI = 5 MW/m2 to COV-IMEPg = 2%.
● Tin = 117°C: CA50 range = 6.5°CA– 0.8° ∆CA50 / 1.0° ∆spark-timing
● Tin = 107°C: CA50 range = 2.4°CA
● CA50 range for acceptable SAcombustion is smaller for lower Tin.
● At these conditions:Flame propagation with SA does not allow CA50 to be more retarded than for CI-mode w/o SA (374° CA).– Pure CI-mode, has virtually the
same CA50 range = 6.4°CA.
● But Spark-Assist gives rapid control.
PM, CF-E0, Pin = 1 bar, φm = 0.42
Collaborations● Project is conducted in close cooperation with U.S. Industry through the
Advanced Engine Combustion (AEC) / HCCI Working Group, under a memorandum of understanding (MOU).– Twelve OEMs, Three energy companies, Six national labs, & Several universities.
● General Motors: Bimonthly internet meetings ⇒ in-depth discussions.– GM provided 300-bar Bosch HDEV5.1 GDI injector and spark-ignition system.– Provide data to GM on boosted LTGC and for modeling PFS-LTGC.
● Cummins, Inc.: Discussions and guidance on working with new low-swirl, spark-plug cylinder heads (Head #2), potential acquisition of Head #3.
● LLNL: Support development and validation of chemical-kinetic mechanism for RD5-87 (87-AKI, E10 gasoline) and related RCM measurements at ANL.
DOE-OVT project is also leveraged through three related research efforts
● Co-Optima Fuels Project: Funds-in project of advanced fuels containing a significant renewable fraction for boosted SI and low-T combustion engines.
● Chevron: Funds-in project on advanced petroleum-based fuels for LTGC.
● Sandia LDRD: Funds-in project on fuel injection.
Future Work● Continue to focus efforts on combustion-timing control & improved robustness,
with an emphasis on lower boost (1.0 ≤ Pin ≤ 2.0 bar).● Use RD5-87 gasoline (regular E10) for now, and reduce CR to 14:1 ⇒ should
increase operating range with RD5-87 and more in-line with OEM targets.– Map engine performance for CR = 14:1 w/ RD5-87 (will reduce TE 1.0 – 1.5 %-units)
DDI-PFS with Variable Inj. Timing: ⇒ CA50 control & multiple other benefits● Determine the range of conditions for which DDI-PFS can be applied effectively
⇒ range of Pin (down to 1.3 bar?), fueling rates (φm), and speed effects.
● Investigate various fueling strategies to improve PFS performance and extend range of application ⇒ vary late-DI timing & fraction, multiple injections, etc.– Image fuel distributions in optical engine to guide strategies.– Potential of 300 bar GDI injector to improve PFS and its operating range.
Spark-Assisted (SA) LTGC: ⇒ CA50 control, etc.● Map out range of conditions for effective SA-LTGC with CR = 14:1.
– Determine benefits at Pin = 1.0 bar, and find max. Pin for effective SA.– Investigate effect of DI fueling and PFS, speed effects, potential to extend load.
Continue to support of LTGC/HCCI modeling: Provide data, analysis, and discussions to support kinetic modeling at LLNL, and CFD modeling at GM.
Summary● A new spark-plug capable, low-swirl cylinder head has been installed, and it’s
combustion performance characterized.– Overall performance is similar to previous head, with two exceptions:
1) For PM fueling, TE is lower by 0.2 – 1.0%-units, due to increased heat transfer.2) For early-DI fueling, TE is also reduced at low and high fueling rates due to
reduced combustion efficiency caused by less complete fuel/air mixing.– High-load limits and CNL are similar for both heads, both PM & DI fueling, all Pins.
● Both CF-E0 & RD5-87 are φ-sensitive for Pins down to at least 1.3 bar, indicating that the benefits of PFS can be obtained ⇒ RD5-87 better at lower Pins.
● Showed injection timing can control CA50 up to 8.6°CA, from strong knock to near misfire, as part of DDI-PFS fueling strategy ⇒ ultra-low NOx & soot.– Retard the late-DI timing increases stratification advances CA50
● Showed that DDI-PFS substantially increases the allowable CA50 range from knock to near misfire. ⇒ It can also increase stability for a significant extension of the load range.
● Spark-Assist was found to be effective for CA50 control & increased Tin tolerancefor φ > 0.36 at Pin = 1 bar. ⇒ Complements DDI-PFS, which works Pin ≥ 1.3 bar.
● Collaborated with LLNL on development of a kinetic mechanism for RD5-87 and supported related RCM measurements at ANL, and with GM on CFD modeling.
Technical Backup Slides
Collaboration: Kinetic Mechanism for RD5-87● RD5-87 is a research-grade 87-AKI, E10 regular gasoline with tightly controlled
specifications. ⇒ Representative of market fuels.● Accurate chemical-kinetic mech. will be valuable for research groups & industry.● Collaborate with LLNL (W. Pitz & M. Mehl) to support their development of a
kinetic mech. for RD5-87, and support related RCM measurements at ANL.● SNL: Engine data recently acquired for RD5-87 for fully premixed operation
over a wide range of Pin and fueling rates (φm).– Data to be provided to LLNL for mechanism tuning and validation.– Provided fuel to ANL for RCM studies.– Discussions with LLNL and feedback on
mechanism performance for further improvement.
● LLNL: Proposed a chemical-kinetic mechanism based on a 5-component surrogate, matching compositional & octane properties. ⇒ will tune and validate based on SNL engine data and ANL’s RCM data as available.
● ANL: RCM data on RD5-87 autoignition.
LLNL proposed surrogate for RD5-87
From W. Pitz, LLNL