Page 1
G. De Sercey, G. J. Awcock and M. HeikalUniversity of Brighton
School of Engineering
[email protected]
This Work Conducted In Association With Ricardo Consulting Engineers, UK
Toward A Calibrated LIF Image Acquisition Technique For In-Cylinder Investigation Of
Air-to-fuel Mixing In Direct Injection Gasoline Engines
OSAV’2004 International Topical Meeting s
Page 2
Toward A Calibrated LIF Image Acquisition Technique For In-Cylinder Investigation Of Air-to-fuel Mixing In
Direct Injection Gasoline Engines
• Introduction
• The Laser Induced Fluorescence (LIF) Technique
• The Optical Set-up for Quantitative Measurement
• Calibration Strategy
• Tracer Optimisation
• Calibration Process
• Conclusion; - Discussion Of Results
Page 3
Introduction IIntroduction IThe Pressure for Better, Cleaner Engines
GDI Engines
User Demand
• Rocketing Fuel Cost
• 1970’s Onwards
• Better Economy Is A Selling Point
• Manufacturers MUST Develop Cleaner Engines To Continue To Sell Cars!
Fuel Injected (PFI) Engines
19831988
1991 (Euro I)1996 (Euro II)
2000 (Euro III)2005 (Euro IV)
CO
HC
NOx
0
10
20
30
40
50
60
70
80
90
100
Percentage of Reduction
Year
Pollutant
European Emission Standards for Gasoline Engines (2.0ℓ)
CO
HC
NOx
Evolution Of EuropeanEmission Standards ForGasoline Engines (2.0 l)
Imposed Pollution Limits
• Widespread Legislation
Page 4
Introduction IIIntroduction IIGasoline Direct Injection Engine: Injection Directly In The Engine Cylinder
• Better Control Over Injection
• Less Heat Losses
• Lower Consumption
• Reduced Emissions
Achieved By Concentrating FuelAround The Spark Plug
• Complex Geometry
• Complex Air Flow
• Complex Air / Fuel Mixing
Stratified Mixture
Intake Port
Spark Plug
Exhaust
Air Flow
Bowl-In-Piston
Injector
Page 5
LIF TechniqueLIF Technique
Ab
so
rpti
on
Flu
ore
sc
en
ce
Qu
en
ch
ing
(l
os
se
s)
ground electronic state
excited electronic state
Rotational vibrational transitions
(Colour shift)
Laser light
Fluorescence
Excited molecule (Tracer)
LI
Em
iss
ion
Page 6
Why Quantitative LIF?Why Quantitative LIF?
Qualitative LIF
Shows Relative Distribution At A Particular Piston Position, Or
Crank Angle (CA)
Quantitative LIF
Shows Absolute Distribution At Any Engine Position
• No Comparison Between Crank Angles
• No Comparison Between Experiments
• Gives Actual Fuel Concentration
• Allows Comparison Between Crank Angles
• Allows Comparison Between Experiments
Page 7
Optical Set-up IOptical Set-up I
Engine withquartz annulus
Laser Nd:YAG, 266nm
Shutter
Sheet forming optics
motor
Beam dump
Schott filter
532nm ‘filter’Lens-coupled gated image
intensifier
CooledCamera
PC
Coated mirror(+ beam monitor
tap)
Page 8
Optical Set-up IIOptical Set-up II
Page 9
Calibration StrategyCalibration StrategyMust Compensate For Dependence Of Fluorescence On T & P
Best Practice So Far: Measure Of T & P Dependency In A Pressure Vessel, BUT…
• Optical Set-up Different From The One Of The Experiment• Unrealistic, As T & P Varies Spatially In The Engine!
In-Cylinder Calibration
• Same Optical Set-up• No Need To Measure P & T Provided Calibration
And Experimental Images Are Acquired At The Same Crank Angle
Page 10
Intake air
Exhaust
Intake plenum
Heating tape
Insulation layer2’’ ID Pipe
Ball valve
Ball valve
Evaporation crucible
Engine
Injection hole
Calibration LoopCalibration Loop
Page 11
Choice of TracerChoice of Tracer
• Absorption Wavelength Achievable With A Laser
• Enough Fluorescence To Be Detectable With Decent SNR
• Low Sensitivity To Quenching
• Similarity To Fuel In Term Of Physical And Vaporisation Properties
• Non-Hazardous!
Characteristics Sought For The Tracer
Page 12
LIF Tracer PossibilitiesLIF Tracer Possibilities
Fuel or Tracer Absorption (nm) Emission (nm) Boiling Point (ºC)
Gasoline 240-300 330 Various
Iso-octane Non fluorescing 99
Biacetyl 300-360 440-480 88
DMA 240-300 335 193
Toluene 240-260 270-370 110
2-Hexanone 240-310 350-450 127
Acetone 240-340 350-450 56
3-Pentanone 240-310 350-450 102
Page 13
Tracer Optimisation ITracer Optimisation I
Test With Pure Acetone Saturation
0
100
200
300
400
500
600
0 0.5 1 1.5 2 2.5 3 3.5 4
Equivalence Ratio (Ø)
Flu
ore
scen
ce I
nte
nsit
y (
a.u
.)
80
100
180
280
300
CrankAngles
What IsEquivalence
Ratio?
Page 14
Tracer Optimisation IITracer Optimisation IITest With Various Acetone Concentrations In Iso-Octane
Optimum Between 2 And 10%
0
200
400
600
800
1000
1200
1400
1600
0%10%20%30%40%50%60%70%80%90%100%
Acetone Concentration
Flu
ore
scen
ce I
nte
nsi
ty (
a.u
)
100CA
180CA
280CA
300CA
Page 15
Calibration Process OverviewCalibration Process OverviewEngine Motored In Closed-Loop Mode
• Calibration Images Acquired (For Each CA And Equivalence Ratio)
And Processed To Extract Average Intensity
• Average Intensities Plotted And Piece-Wise Linear Fitted
• Calibration Look-Up-Table (LUT) Generated
Engine Motored In Normal Mode
• Fuel Mixing Experiments Performed & Images Acquired
• (Error Images Derived, At Each CA, Mid-Term, BUT In Closed Loop Mode)
• Error Image Corresponding To The Same CA Subtracted
• Calibration Map Applied
Quantitative Air-to-Fuel Ratio Maps
Page 16
Calibration Process SummaryCalibration Process Summary
Error Subtraction
Calibration
Quantitative Data
Raw Experiment Image
-
Error Image
Corrected Image
Page 17
Review; - Why Quantitative LIF?Review; - Why Quantitative LIF?
Qualitative LIF
Shows Relative Distribution At A Particular Crank Angle
Quantitative LIF
Shows Absolute Distribution At Any Engine Position
• No Comparison Between Crank Angles
• No Comparison Between Experiments
• Gives Actual Fuel Concentration
• Allows Comparison Between Crank Angles
• Allows Comparison Between Experiments
Page 18
Quantitative Results IQuantitative Results I
Equivalence Ratio Scale:
Page 19
Quantitative Results IIQuantitative Results II
0
100
200
300
400
500
600
700
800
50 100 150 200 250 300 350
CA
aver
age
flu
ore
scen
ce (
a.u
.)
1500-.5-61
1500-30-61
1500-60-61
1000-60-45
500-60-20
Average equivalence ratio
0
0.5
1
1.5
2
2.5
3
3.5
50 100 150 200 250 300 350
CA
equ
ival
ence
rat
io 1500-.5-61
1500-30-61
1500-60-61
1000-60-45
500-60-20
Uncalibrated Fluorescence Calibrated Fluorescence; - Equivalence Ratios
Crank-Angle Compensation Allows Valid Fuel Mixing Studies To Be Conducted Over All Relevant Crank Angles
• A Range of Injection Strategies (At 1500 RPM)• Start of Injection (SoI) At 0.5º, 30º, 60º ATDC
• A Range of Engine Speeds (At SoI 60º ATDC)• 1500, 1000, 500 RPM
Page 20
Review; - Why Quantitative LIF?Review; - Why Quantitative LIF?
Qualitative LIF
Shows Relative Distribution At A Particular Crank Angle
Quantitative LIF
Shows Absolute Distribution At Any Engine Position
• No Comparison Between Crank Angles
• No Comparison Between Experiments
• Gives Actual Fuel Concentration
• Allows Comparison Between Crank Angles
• Allows Comparison Between Experiments
Page 21
Quantitative Results IIIQuantitative Results III
Mixture Distribution at 90º CA For A SoI At TDC, With Superimposed DFVR Air-Flow Predictions
Comparison With Dynamic Flow Visualisation Rig (DFVR)• DFVR Is A PIV Technique
Using Water Seeded With Particles To Visualise Flow
• LIF And DFVR Results Are Compared At The SAME Crank Angle
– Good Correspondence
• Rich Mixture (1.2<Φ<1.8) On Exhaust Side
– Carried With Flow Out Of Bowl
• Lean (Φ<0.5) On Intake Side• Dilution By High Velocity Air From
Open Intake Valve
Page 22
Quantitative Results VIQuantitative Results VICoefficient of Variation (CoV) Can Be Determined To Study Stability Of The Mixing Process
• CoV Is The Image RMS Difference Values Divided By Image Mean
CoV Mixture Stability for Various Start of Injection Timings (White = >25%)
Injection at TDC Injection at 30CA Injection at 60CA25%
10%
0%
• These Results Suggest That 30CA Is The Most Stable Scenario
• Tests Performed On A Firing Engine Support This Evidence
• Injection At 30CA Gives Best Emissions Performance And Minimum ‘Knock’ (Pre-ignition)
Page 23
ConclusionsConclusions
• A New Strategy Has Been Developed For Calibration of LIF Measurements
− Critical To Understanding Air-Fuel Mixing In The Cylinder
• It Is Efficient And Realistic− Thanks To Calibration At Full Range Of Equivalence
Ratios, Crank Angles And Engine Speeds
• It Is Effective− Predictions From Motored Test Engine Give Good
Agreement With Independent Investigations