GASOLINE COMPRESSION IGNITION – A PROMISING …

GASOLINE COMPRESSION IGNITION – A PROMISING TECHNOLOGY TO MEET FUTURE ENGINE EFFICIENCY AND EMISSIONS TARGETS

drhgfdjhngngfmhgmghmghjmghfmf

STEPHEN CIATTIPrincipal Mechanical EngineerArgonne National Laboratory

2nd CRC Advanced Fuels and Engine Efficiency WorkshopWednesday, November 2, 2016Work funded by DOE Office of Vehicle Technologies – Gurpreet Singh and Leo Breton

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

KHANH CUNGPostdoctoral FellowArgonne National Laboratory

INTRODUCTION Advanced combustion concept: HCCI, PPC

or GCI utilizing LTC process to reduce soot formation, and avoid high NOx (more premixing and dilution) GCI has shown to have advantages over

traditional diesel compression ignition due to: low soot and emission, high thermal efficiency Initial hardware tends to influence desired

strategies and tradeoffs Many dependent variables (Pin, Tintk, mixture

stratification) on the ignition, combustion and the emissionRecently explored: Effect of multiple parameters (injection pressure, mixture composition, boost,

injection timing, and fuel reactivity) on engine outputs (ignition, combustion phasing, emission, etc.) using design of experiment approach

Distinguish the transition from “quasi” HCCI to GCI as injection timing is swept with constant air/fuel mixture for two different fuels

Kitamura, T., et al. International Journal ofEngine Research 3.4 (2002): 223-248.

ENGINE SPECIFICATION4-cylinder light duty diesel engine (Euro IV

General Motors 1.9 L)Diesel fuel injection system (Bosch CRIP2)Eaton supercharger equipped for study with

stable boost pressure

1234

Supercharger

Uncooled EGR Loop

Cooled EGR Loop

Heat Exchanger

Exhaust Outlet

Ambient Intake Air (from LFE)

Bypass Valve

Heat Exchanger

Intake Air (+ EGR)

Exhaust

Turbocharger

Engine GM 1.9 L (4 cylinders)

Bore 82 mmStroke 90.4 mmConnecting rod 145.4 mmCompression ratio 17.8:1IVO 359 deg. aTDCIVC -149 deg. aTDCEVO 130 deg. aTDCEVC -357 deg. aTDCEngine speed 1000 RPM

2000 RPMIntake pressure (abs.)

1.15 to 1.55 bar

Intake temperature 48 deg. CInjectors Bosch

Numbers of hole 7Nozzle diameter 0.141 mmSpray angle 120 deg.

Injection pressure 400 to 600 bar

Graphics courtesy ORNL (Curran & Dempsey)

OBJECTIVES OF WORK - MULTI-CYLINDER, HIGH EFFICIENCY GASOLINE COMPRESSION IGNITION

Current Specific Objectives:1. Evaluate effect of Low Pressure EGR upon auto-ignition

and engine performance characteristics 2. Quantitatively study effect of injection strategy upon auto-

ignition to develop approach for transient operation and reduced fuel sensitivity

3. Perform factorial experiments to quantify the effect of important input parameters upon engine performance, noise and emissions

4

Long-Term ObjectiveUnderstand the physical and chemistry characteristics ofGasoline Compression Ignition (GCI) in a multi-cylinder engine toaid industry in developing a practical high efficiency, lowemission combustion system

Mixing Limited GCIHCCI PFS

Majority Premixed GCI

Majority Stratified GCI

INJECTION STRATEGY: E10 MINIMUM FUELING – FULL SOI SWEEP (-141 ATDC TO TDC)

5

SOI [deg. aTDC]

-140 -120 -100 -80 -60 -40 -20 0

Lam

bda

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Lambda

Minimum Fueling approach: least fuel requirement for stable combustion (COVIMEP < 3%)

Combustion mode (HCCI vs. GCI) characterized by SOIs

Discontinuity due to fuel rate changed

A-period Location of ignition (CA10), and combustion phasing (CA50) seems stay constant

B-period More fuel (smaller lambda) is needed to have stable combustion, but CA10/CA50 also seems constant

C-period IMEP shows a drop near -60 deg. aTDC due to possible fuel entering squish region, as indicated by simulation

A B C

Fuel in squish region

Both SOI and Lambda show to have effect on CA10/CA50, but lambda is more effective

There seems to be a condition with constant lambda to fix CA10/CA50

GM 1.9 L 17.8:1 (CR)Engine speed 1000 rpm

Injection pressure 400 bar

Injector-Bosch

7 hole 120 deg.

cone angleFuel E10

Constant lambda at some SOI range

SOI [deg. aTDC]

-140 -120 -100 -80 -60 -40 -20 0C

A [d

eg. a

TDC

]-10

-5

0

5

10

15

20 CA10CA50CA90

E10 CONSTANT LAMBDA – SOI SWEEP

Early to near TDC SOI effect

Minimum fuel provides fuel requirement (least fuel) for combustion stability, but does not give same mixture to study SOI effect explicitly on ignition and combustion constant lambda approach

Fuel rate was adjusted to keep same lambda through all SOIs

λ calculated from emission bench

“Sweet spot” (early ignition)

High noise level near

TDC

quasi HCCI Transition GCI

Less fuel in squish region

Almost similar ignition location in “quasi HCCI” (also similar CA50/CA90)

Existing region where earliest ignition occurs (near -30 deg. aTDC) – reduction in fuel in squish region

Near TDC, short residence time for ignition

IMEP increases slightly near TDC (less fuel in squish)

It was harder to control noise level with late injection

very lean combustion (λ = 4.4)

6

COVIMEP<3%, max = 5%)

LOW SPEED/LOAD E10 CONSTANT LAMBDA – SOI SWEEP: EMISSION

Early injection: Low level of NOx in highly

homogeneous mixture (quasi HCCI region)

Incomplete combustion (high HC)

Low combustion temperature high CO

Smoke number (FSN) was very low (<0.1) due to lean condition at all SOIs

Late injection (GCI mode): Insufficient time for air/fuel mixing High NOx due ~ high combustion

temperature; richer zones for ignition

Less HC and CO (aggressive reaction leads to more complete combustion)



Injector-Bosch

7 hole 120 deg cone

angleFuel E10

7

SOI [deg. aTDC]

-140 -120 -100 -80 -60 -40 -20 0H

C[g

/kw

-hr]

0

2

4

6

8

10

HC

SOI [deg. aTDC]

-140 -120 -100 -80 -60 -40 -20 0

T ex

haus

t [de

g. C

]

110

120

130

140

150

160

170

NO

x [g

/kw

-hr]

0246810121416

TexhaustNOx

SOI [deg. aTDC]

-140 -120 -100 -80 -60 -40 -20 0

CO

[g/k

w-h

r]

0

20

40

60

80

100

120

CO

Low combustion temperature

More fuel enters squish zoneLong residence

time, well-mixed combustion

PARAMETRIC STUDY: HIGHER ENGINE SPEED CONDITION (2000 RPM)

Engine Speed 2000 rpmLevel Low High

Constant Boost 0.45Injection pressure [bar] 400 600

SOI [deg. BTDC] 70/20 70/40Lambda 2.7 3.7

Constant Lambda 3.1Injection pressure [bar] 400 600

SOI [deg. BTDC] 70/20 70/40Boost [bar] 0.35 0.55

Double injection: Same duration for

pilot & main Fixed pilot Helpful for meeting

COV, noise levels

Sample AHRR of long vs short dwell: Pinj=600 bar, PIntk=0.55 bar, λ = 3.1

8

Considered parameters

P Injection PressureS Start of injectionL LambdaB Boost

FACTORIAL STUDY - CONSTANT BOOST TESTS SHOW INFLUENCE OF COMBUSTION MODE ON EMISSIONS, LAMBDA ON COMBUSTION PHASING

EffectP S L PS PL SL PSL

Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5

IMEPCOVNoiseBSFC

EffectP S L PS PL SL PSL

Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5 NOx

HCCO

B [bar] 0.2P [bar] 400 600S [deg. bTDC] 15 141L 3.6 4.5


1000 RPM

2000 RPM

λ has strong impact on ignition

Lower performance with leaner mixture

Lower emissions with shorter dwell

B [bar] 0.45P [bar] 400 600S [deg. bTDC] 70/20 70/40L 2.7 3.7

FACTORIAL STUDY - CONSTANT LAMBDA TESTS SHOWS SIGNIFICANT BOOST EFFECT ON COMBUSTION PHASING, NOISE AND EMISSIONS

10

EffectP S B PS PB SB PSB

Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5 CA10

CA50CA90


Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5 NOx

HCCO


Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5 CA10

CA50CA90


Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5

IMEPCOVNoiseBSFC


Nor

mal

ized

[a.u

.]

-1.5-1.0-0.50.00.51.01.5

NOxHCCO

1000 RPM

2000 RPM

Advanced ignition at higher boost

Reduced COV, BSFC, Increased Noise at higher boost

Overmixed at high P_inj(high HC, CO)


L 4.2P [bar] 400 600S [deg. bTDC] 15 141B [bar] 0.15 0.3

L 3.1P [bar] 400 600S [deg. bTDC] 70/20 70/40B [bar] 0.35 0.55

0

20

40

60

80

100

120

140

0

5

10

15

20

25

30

‐150 ‐100 ‐50 0 50 100 150

HRR

[J/CAD

]

Curren

t [A]

CA [aTDC]

Injector Current and HRR of 5 bar and 8 bar at 2000 RPM with ~30% EGR (NOx<0.45 g/kWhr, CO<1 g/kWhr, HC<3g/kWhr, FSN < 0.03)

20160404_000

20160404_001

20160404_002

20160405_002

20160405_003

HRR_0404_000

HRR_0404_001

HRR_0404_002

HRR_0405_002

HRR_0405_003

SOI OF MULTIPLE INJECTION AND HRR: 5 BAR AND 8 BAR BMEP

11

SO

I 1st

= -1

00 a

TDC

SO

I 2nd

= -7

0 aT

DC

SO

I 3rd

= -2

5 aT

DC

Pinj = 600 bar5 bar

(48 C, 1.37 bar)

8 bar(48 C, 1.63

bar)

CAD

0 10 20 30 40 50

HR

R [J

/CA

D]

-20

0

20

40

60

80

100800 bar SOI (47/29/16)600 bar SOI (100/70/25)

INJECTION PRESSURE & SOI EFFECTIVELY CONTROL COMBUSTION PHASING

12

CA10/50/90 delayed by 4

deg

Retarded SOI further delay ignition and combustion phasing: this results in slight reduction of noise, but most of all relax the constraint of EGR

Higher peak pressure near TDC for late SOI case implies higher heat loss. This resulted in lower performance of BSFC (increase by 22 g/kW-hr). Adjustment in 1st

and 2nd injection and/or increase in EGR would possibly help

CHALLENGE IN FUNDAMENTAL STUDY OF SOI EFFECT

13

Fueling rate varied when changing SOI Injector per cylinder needs to adjusted SOI duration to

match load, combustion phasing

Slight change in 3rd

injection timing (0.2 deg) could make a difference in noise level (~4 dB level)

CAD

-5 0 5 10 15 20 25 30 35N

orm

aliz

ed M

FB

0.00.10.20.30.40.50.60.70.80.91.01.1

Cyl 1Cyl 2Cyl 3Cyl 4

LP EGR SETUP & TESTING CONDITIONLP EGR Adjustment by means

ExhaustValve

Throttle valve on overall exhaust discharge

Inlet ValveThrottle on fresh intake air upstream of turbo, to drive LP-EGR

LP EGR Valve

Throttle valve between post DPF exhaust and turbocharger intake

Test condition: EGR% sweep at constant load

(BMEP ~ 3 bar at 2000 RPM) EGR% adjusted by separate valves

(most effective exhaust valve) Triple injection (SOI of each: 100-70-

25 deg. bTDC) Supercharger (ON) for P intake = 0.6

bar• Allows for precise control of

intake pressure!

Inj 1 Inj 2 Inj 3

Pinj = 400 bar

14

Smoke meter

DPF

AMA2000

Lambda sensor

RECENT RESULTS WITH LP-EGR: APPROACHING AN OPTIMAL INJECTION STRATEGY FOR LOW NOISE, BSFC AT FULL LOAD RANGE

15

S/C OFF (new test) Since start of injection is very in

controlling the combustion phasing; noise is also reduced

EGR was even reduced from 30% to 21% even for late injection

Higher EGR may be required to maintain location of combustion phasing

Advantage of using higher injection pressure for fuel atomization and promote mixing to achieve lean combustion (lambda ~ 1.6-1.9); more stable combustion

Speed Rail P EGR P_intk T_intk Lambda BMEP BSFC CA10 CA50 CA90 IMEP COV Noise NOX HC CO FSN T_exhaust[rpm] [bar] [%] [bar] [deg. C] [a.u.] [bar] [g/kw‐hr] [bar] [%] [dB] [a.u.] [deg. C]2000 600 29.8 1.4 45.6 1.4 4.9 312.0 7.8 10.8 17.6 6.1 3.9 93.9 0.4 1.2 3.2 0.017 316.92000 800 21.1 1.5 48.1 1.6 4.8 337.5 12.4 16.1 25.7 6.0 3.3 91.8 0.5 0.1 0.4 0.024 452.8

[g/kw‐hr][aTDC]

Low emission

LP-EGR IMPROVES PERFORMANCE AT HIGHER LOADS AS WELL

Emissions (NOx, HC, CO) and FSN are reduced significantly with LP- EGR

Exhaust temperature remains high with EGR• Improved aftertreatment expected with higher

exhaust T Supercharger needed to maintain boost stability

• Higher BSFC than intended• LP-EGR modifications to allow use of turbo

only



Injector-Bosch

7 hole 120 degcone angle

Fuel E1016

EGR Intake T T_exh BMEP BSFC Noise NOX HC CO FSN[%] [deg. C] [deg C] [bar] [g/kw-hr] [dB] [g/kW-hr] [g/kW-hr] [g/kW-hr] [a.u.]

29.41 53.64 309.5 4.32 346.13 90.13 0.45 0.91 2.10 0.03029.84 45.56 316.9 4.91 312.04 93.90 0.42 1.16 3.20 0.01729.43 47.11 327.5 5.16 317.10 91.58 0.33 0.57 1.34 0.02030.50 47.58 419.9 8.27 280.25 93.88 0.04 0.37 1.27 0.02730.48 47.91 412.2 8.33 278.78 91.01 0.05 0.37 1.44 0.024

17

Study at UCB (Vuillemier) shows that LTHR has significant effect on gasoline HCCI ignition

His conclusions indicate:

A fuel’s Octane Index is a good indicator of its GCI Low-Load Performance

LTHR Onset Pressure in an HCCI engine correlates very well with GCI Low-Load Performance.

Increased intake pressure increases low-temperature heat release, enabling lower loads in a GCI engine

COLLABORATION - UCB WORK SHOWS SIGNIFICANT LTHR DEPENDENCE FOR GCI

Courtesy of David Vuilleumier (August 2015 AEC meeting)

SUMMARY New approach to study the ignition of E-10 in GCI engine using “constant lambda” to

investigate ignition behavior Combustion modes were seen distinctively from CA10/50/90 vs SOI plot: “quasi”

HCCI, transition, and GCI modes Design of experiment (DOE) provide overall view of different parametric study using

minimal number of test runs: constant lambda, constant boost– Boost, lambda, and SOI affect strongly on ignition, combustion phasing, and emission– Injection pressure has minimal effect at low speed; but contribute significantly at high speed

in controlling NOx/HC/CO, and combustion stability

Installation of the new LP-EGR has been evaluated with benefit of using smoke-filtered and cooled EGR to effectively control the combustion phasing, therefore noise and emission– Higher EGR decrease combustion noise but fuel consumption is still high due to incomplete

combustion– Reduced EGR required for 5 bar BMEP was achieved by retarding SOIs. This was

performed with higher injection pressure with the same emission output (low emission, and ultra-low smoke)

– Combustion phasing was further delayed to facilitate with combustion noise– The SOI of the 3rd injection (determine load and combustion phasing the most) is very

sensitive to both noise and BSFC. Further investigation is needed to identify the flexibility when changing higher load/speed

18

FUTURE WORK Continue to explore/understand effect of injection strategy

upon GCI operation E10 is sensitive to both boost and EGR

Explore more conditions with LP-EGR Provide more boost at low speeds/loads with EGR

Examine influence of CR upon combustion noise Alter IVC relative to exhaust cam Use lower CR piston crowns (we have 16, 15 and 14:1)

Continue to develop strategy for transient operation with injection, boost and EGR

Continue to track and account for USCAR guidelines combustion noise – later injections at higher pressure show significant promise! Target <90 dB for high load, <85 dB for low load

Continue to characterize GCI particulate emissions TEM sampling and analysis19

www.anl.gov

THANK YOU FOR YOUR ATTENTION!

QUESTIONS?

GASOLINE COMPRESSION IGNITION – A PROMISING …

Documents