AB, 16 08 04 1

On-line diagnostics and fast modeling of sootformation / oxidation in diesel engine

combustion

A Bertola, S Kunte, M Warth, P Obrecht and K Boulouchos

Aerothermochemistry and Combustion Systems Laboratory

Swiss Federal Institute of Technology

ETH Zürich

8 ETH Conference on combustion generated nanoparticlesAugust 16, 2004

AB, 16 08 04 2

At the Aerothermochemistry and Combustion Systems Laboratory of the Swiss Federal Institute of Technology in Zurich we are currently developing low emissions strategies for heavy-duty diesel engines that engine manufacturers can implement to meet stringent emissions regulations. The technologies being studied include high-pressure fuel injection (with common-rail injection system), turbocharging, and exhaust gas recirculation (cooled EGR). We also investigate the influence of oxygenated diesel fuel additives and water-diesel fuel emulsions on combustion and emissions [1].Measurements are carried out on a heavy-duty four-cylinder diesel engine equipped with turbocharger and common-rail fuel injection. Wide variations of injection parameter settings and EGR rate are performed. Engine experiments are conducted with reference diesel fuel and 3 different water-diesel fuel emulsions (13% / 21% / 30% Water) and with a blend of reference diesel fuel and butylal.The experimental study includes measurements of pollutants in the exhaust gas (Opacity, Filter Smoke Number, Particulate Number Size Concentration, gaseous components as NOx, HC ...) as well as in-cylinder on-line measurements of soot concentration / soot temperature by means of the two-color pyrometry method.

An optical probe developed by Schubiger in 2001 [2] is used to collect the light intensity of the soot radiation during combustion. The soot concentration (KL-factor) and the transient burned gas temperature are computed with the multicolour-pyrometry technique (3 wavelengths). The measurements show that the KL-signal start coincides with the beginning of the mixing-controlled combustion phase where soot is first formed. The soot formation and oxidation processes are clearly recognizable form the KL-data plots over time. The measured transient burned gas temperature reaches a maximum at the occurrence of “injection process ends”. A match of the measured burned gas temperature with a computed temperature in the combustion chamber obtained by varying the air/fuel ratio (A/F) in the 2-zone model, shows that the local A/F during soot formation starts at 0.55 (fuel rich) and ends up at ca. 0.9 (near stoichiometric) at the end of the soot oxidation phase. The variation of the fuel composition (use of water-diesel fuel emulsions or oxygenated diesel additives) leads to reduced soot formation rate and enhanced soot oxidation rate. Moreover, at higher injection pressures there is clear indication that the soot oxidation process begins earlier and is more effective. The measured soot radiation temperature is lower when using water-diesel fuel emulsions respectively abot the same as that of the reference diesel fuel when using oxygenated diesel additives. The exhaust gas recirculation lowers the soot oxidation process dramatically, this fact is responsible for having higher particulate emissions at higher EGR rates.Data of in-cylinder soot concentration at the end of the soot oxidation phase (KL-factor after the drop passed the second KL-maximum) correlate clearly with particulate measurements in the exhaust (total number of particles 20-500 nm and filter smoke number) [3].

AB, 16 08 04 3

The LAV approach for fast modeling of soot formation / oxidation processes during the combustion is accomplished with phenomenological models (0-D fast modeling) and with 3D-simulations. It is based on a simplified and essentially correct description of the physical and chemical mechanisms prevailing in diesel engine combustion. The approach includes submodels for the prediction of heat release rate, NO and soot emissions [2, 4]. The model parameters (ca. 10 for the soot model) are identified and optimized with bioinspired evolutionary algorithms (stochastic, parallel search method) using typically 20 engine measurements with a wide range of operating conditions [4]. The model validation carried out for different engine operating conditions shows very promising results; the model can predict the particulate emissions of diesel engines with high accuracy and in short time. It is for example possible to compute 1Mio operating conditions in the engine map within 2 days using 3 PC’s, while for the same task the optimization of PM emissions at the engine test bench would take 1000 times more.

Conclusions and outlook:• Particle number size measurements in the exhaust gas have been combined with relevant combustion parameters and in-cylinder measurements of soot concentration.• The combination of a flexible -high pressure- fuel injection system, exhaust gas recirculation, oxygenated fuels and water-diesel fuel emulsion technology allowed to separate important thermochemical and fluidmechanical influences.• Soot measurements in the combustion chamber and in the exhaust were used for the development and the validation of phenomenological soot models which can help to optimize the diesel engine combustion process in shorter time.

References:[1] Bertola A : Technologies for Lowest NOx and Particulate Emissions in DI-Diesel Engine Combustion - Influence of Injection Parameters, EGR and Fuel Composition, Thesis dissertation ETH Zurich nr.15373, 2003.[2] Schubiger S : Untersuchungen zur Russbildung und -oxidation in der dieselmotorischen Verbrennung: ThermodynamischeKenngrössen, Verbrennungsanalyse und Mehrfarbenendoskopie, Thesis dissertation ETH Zurich nr.14445, 2002.[3] Bertola A et al : Temporal Soot Evolution in Diesel Engine Combustion – Influence of Fuel Composition, Injection Parameters and EGR, 2004 in preparation.[4] Warth M et al: Prediction of Heat Release Rates, NO- and Soot Emissions in Diesel-Engines – Optimization and Validation of a New Approach, 9th Symposium "The Working Process of the Internal Combustion Engine", Institute for Internal Combustion Engines and Thermodynamics, Graz University of Technology, 2003.

AB, 16 08 04 4

Research approach & motivation

Heavy-duty diesel enginecommon rail injection system

Fuel-related parametersFuel injectionFuel composition

Air-related parametersExhaust gas recirculationTurbocharging

Combustion analysis & Modeling of soot formation/oxidation

Exhaust emissions

Crank angle-resolved data ofin-cylinder soot concentration

modeled

AB, 16 08 04 5

Heavy-Duty Diesel EngineLIEBHERR 4-cylinder 4-stroke direct injected diesel engine

Stroke = 142 mm; Bore = 122 mmVe = 6.64 l ; ε = 17.2183 kW @ 2100 1/min1060 Nm @ 1540 1/min

Common Rail Injection SystemETH pump (-2000 bar)BSG electronic (main, pilot & post injection)CRT injectors (-1600 bar), Type P2Bosch injectors (-1200 bar)Nozzle tips 6*0.210, 8*0.200 mm

TurbochargerCompressor K 27.2 (original)Turbine casings T21, T15, T12

EGR-System (preliminary)Cooled EGR (high pressure side)With throttle after turbine

AB, 16 08 04 6

Fuel properties

11.98

20.6

15.36

10.15

0

[%]

Oxygen ContentmO2/mtot

0

30

21

13

0

[%]

Water contentmW/(mD+mA)

12.57

11.77

12.38

12.98

14.64

[-]

Stoichiometricair/fuel ratio

82938.2560% butylal blend

87132.5430% W-D emulsion

[kg/m3][MJ/kg]

86234.9621% W-D emulsion

85137.9813% W-D emulsion

81943.14Referencediesel

DensityLower heating

value

AB, 16 08 04 7

Diagnostics on a HD diesel engineOptical in-situ soot observation; multi-colour pyrometry

Spray pattern 8-hole nozzleSpray pattern 6-hole nozzle

Access 1optical

Access 2pressure

Sensor position(field of view, „measuring volume“)

Sensor position within thecylinder head

Source: R Schubiger 2001

AB, 16 08 04 8

Optical lightwaveguide diagnostics

Source: R Schubiger 2001

AB, 16 08 04 9

On-line, global information onsoot concentration and (radiation) temperature

Reference diesel fuel, 1250 rpm, 10 bar BMEP, p rail 800 bar, SOI 1°CA ATDC

360 370 380 390 400 410 420 430

1000

1500

2000

2500 T Pyrometry signal 1 T Pyrometry signal 2 T Pyrometry signal 3 T Unburned air T Burned gas (A/F=global) T Ad. flame (A/F=1) A/F variable T A/F variable

Tem

pera

ture

[K]

Crank angle degrees

0

2x10-9

4x10-9

6x10-9

KL1 KL3 KL5

KL-

Fact

or [-

]

0.550.600.650.700.750.800.850.900.951.00

A/F

rat

io [-

]

360 370 380 390 400 410 420 4300

1

2

3

4

5

6

Bur

n R

ate

[%/°

CA

]

Injection rate

AB, 16 08 04 10

0

100

200

300

Start of InjectionEnd o f Injection

Ref. Diesel, pRail 800 bar 13% Water Em., pRail 800 bar 30% Water Em., pRail 800 bar 60% Butylal Bl., pRail 800 bar

Constant Injection Pressure

Bur

n R

ate

dQ

/ d φ

[J/

°CA

]

0

2x10-9

4x10-9

6x10-9

KL

-Fac

tor

[-]

360 370 380 390 400 410 420 4301600

1800

2000

2200

2400

Crank angle degrees

Mea

sure

dT

empe

ratu

re [

K]

Multi-colour pyrometry variation of fuel composition and injection parameters

0

100

200

300

Start of InjectionEnd o f Injection

Ref. Diesel, pRail 800 bar 13% Water Em., pRail 940 bar 30% Water Em., pRail 1060 bar 60% Butylal Bl., pRail 940 bar

Constant Injection Duration

Bur

n R

ate

dQ

/ d φ

[J/

°CA

]

0

2x10-9

4x10-9

6x10-9

KL

-Fac

tor

[-]

360 370 380 390 400 410 420 4301600

1800

2000

2200

2400

Crank angle degrees

Mea

sure

dT

empe

ratu

re [K

]

AB, 16 08 04 11

Multi-colour pyrometry Variation of fuel composition and EGR rate

0

100

200

300

400

Start of InjectionEnd of Injection

Ref. Diesel, no EGR Ref. Diesel, 12.6% EGR 30% Water Em., no EGR 30% Water Em., 12.6% EGR

Constant Injection Duration

Bur

n R

ate

dQ

/ dφ

[J/

°CA

]

0

2x10-9

4x10-9

6x10-9

8x10-9

1x10-8

KL-

Fact

or

[-]

360 370 380 390 400 410 420 4301600

1800

2000

2200

2400

Crank angle degrees

Mea

sure

dTe

mpe

ratu

re [K

]

AB, 16 08 04 12

8 10 100 5000.0

5.0x1013

1.0x1014

1.5x1014

2.0x1014

0 2 4 60.00

0.02

0.04

0.06

0.08

0.10 Diesel 21% water 30% water 60% butylal

Par

ticle

con

cen

trat

ion

[1/

kWh

]

Particle diameter dp [nm]

PM

[g/k

Wh

]fr

om

NS

D

NOx [g/kWh]

Results: influence of fuel compositionEngine speed 1250 rpm, 50% loadInjection strategy DInjection pressure: 800 bar (diesel) - 970 (21% em) - 1050 bar (30% em) -

930 bar (60% butylal)

AB, 16 08 04 13

„ETH-LAV“ approach – soot modeling

FundamentalsBoulouchos et al. (2001)1

1 Schubiger, R. et al. Russbildung und Oxidation bei der dieselmotorischen Verbrennung.MTZ 5/2002 (63), pp. 342-353.

3 Basic Equations (Formation, Oxidation & Sum-up)Dimensional Correct FormulationOxidation ~ Inverse Characteristic Mixing-Time

Akihama et al. (2001)2

2 Akihama, K. et al. Mechanism of the smokeless rich diesel combustion by reducing temperature. SAE 2001-01-0655, Warrendale, PA, USA, 2001.

„Φ-T diagram“ for soot formationBasis: CHEMKIN „stirred reactor“ calculations

(including PAH formation, soot particle nucleation, coagulation, growth and surface reactions)

Source: Akihama et al.

Based on detailed CHEMKIN calculations

AB, 16 08 04 14

„ETH-LAV“ approach – concept

Settings„Basic“ Principles of Physics & ChemistryAim: Computing Times « 3D-CRFD

Accuracy » Empirical Models

ApproachPhenomenological Models

Dimension adjusted EquationsElementary Physics & Chemistry included

Parameterization Evolutionary AlgorithmsNo Dependency on Operating Conditions (!)

Validation / ApplicationDifferent Engine Set-Ups

(Vd = 0.5 .. 1000 [l/cyl])

1Fuel DiffFuel Evap

Diff

dmC mdt τ= ⋅ ⋅

Chemistry Physics

Efficient Pheno-menological Models

Selection

Recombination

Mutation Re-Insertion

Fitness Evaluation

(Diffusion Flame)

(Parameter Optimization with Evolutionary Algorithms)

AB, 16 08 04 15

Example – soot emissions – parameter identification

accuracy of measurements

Simulation Experiment

PM

[g

/kW

h]

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

Operating Conditions [-]2 4 6 8 10 12 14 16 18 20

Range ofOp. Conditions

Engine4-cyl. CR-DI Diesel(Vd ≈ 1.6 [l])

n1250 .. 1830 [min-1]

pme3.7 .. 14 [bar]

pinj400 .. 1400 [bar]

SOI-14 .. 0 [°CA aTDC]

EGR0 .. 30 [%]

AB, 16 08 04 16

Example – soot emissions – model validation

Simulation Experiment

PM

[g

/kW

h]

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

Operating Conditions [-]22 24 26 28 30 32 34 36 38 40

Range ofOp. Conditions

Engine4-cyl. CR-DI Diesel(Vd ≈ 1.6 [l])

n1250 .. 1830 [min-1]

pme4.0 .. 16 [bar]

pinj350 .. 1600 [bar]

SOI-12 .. 3 [°CAaTDC]

EGR0 .. 43 [%]

accuracy of measurements

AB, 16 08 04 17

Conclusions and outlook

Particle number size measurements in the exhaust gas have been combined with relevant combustion parameters and in-cylinder measurements of soot concentration

The combination of a flexible -high pressure- fuel injection system, exhaust gas recirculation, oxygenated fuels and water-diesel fuel emulsion technology allowed to separate important thermochemical and fluidmechanical influences

Soot measurements in the combustion chamber and in the exhaust were used for the development and the validation of phenomenological soot models which can help to optimize the diesel engine combustion process in shorter time

AB, 16 08 04 18

Acknowledgements

M. Mohr and U. Mathis (EMPA)

Swiss Federal Office of Energy (BFE)

LIEBHERR Machines Bulle S.A.

Common Rail Technologies AG

Lubrizol Corporation

Lambiotte & Cie S.A.