ICE Workshop, STAR Global Conference 2012 March 19-21 2012, Amsterdam Simulation of the Mixture Preparation for an SI Engine using Multi-Component Fuels Michael Heiss, Thomas Lauer
Content Introduction & Motivation Mesh Description Definition and Implementation of Multi-Component Fuels Verification of Spray Dynamics Analysis of Piston Cooling due to Wall Wetting Wall Film Formation for Single- and Multi-Component Fuels Conclusion Summary and Outlook Michael Heiss Sheet 2
Introduction In order to reduce soot emissions and oil dilution of future SI-engines with direct injection a fundamental knowledge about wall film formation and its prediction is crucial. In the presented work, a turbocharged SI engine with direct injection was used to analyze the influence of different injection timings on the wall film formation. Numerical investigations and accompanying experiments at the engine test bench have been carried out to show the benefit of the multi-component fuel approach. Michael Heiss Sheet 3
Motivation high abrupt increase of soot formation 20 CA low early late The measured soot increased abruptly after a critical threshold of the injection start timing (SOI) was exceeded. Soot is an indicator for a diffusive combustion of wall film Analysis of the differences in wall film formation depending on SOI Michael Heiss Sheet 4
Mesh Description Mapped meshing in es-ice V4.16 6-hole injector detailed modelling of the spark plug geometry thin boundary layers Michael Heiss Sheet 5
Fuel Definition RON95 I 7 component approach having the same distillation curve as gasoline reference fuel high boiling temperature low-boiling temperature Batteh, J. J.; Curtis, E. W.: Modeling Transient Fuel Effects with Alternative Fuels, SAE Paper 2005-01-1127 Michael Heiss Sheet 7
Fuel Definition RON95 II Michael Heiss Sheet 8
Fuel Definition RON95 III Implementation in user subroutine dropro.f Bulk Properties: Calculated for the 7 component mixture (density, viscosity, surface tension coefficient, thermal conductivity) Component properties: Vapour pressure for each component as temperature dependant functions having the form of the Clausius Clapeyron equation: y = y 0 + C 1 e T C 2 Michael Heiss Sheet 9
Spray Dynamics Verification with Optical Measurements 1 CA after injection start 10 CA after injection start 25 mm 50 mm Droplet spectrum definition according to PDA spray measurements Good correlation of spray penetration Michael Heiss Sheet 10
Analysis of Piston Crown Cooling due to Wall Film Wetting Motivation: Checking if a constant piston wall temperature is acceptable Static Mesh @ 450 CA A moving mesh with solid cells is not possible yet. Therefore, the analysis was performed on a static mesh. Calculation of the piston temperature considering evaporative cooling and heat conduction Constant Ambient Temperature Fluid Cells Solid Piston Cells (Aluminium) Michael Heiss Sheet 11
Analysis of Piston Crown Cooling due to Wall Film Wetting Formation of Liquid Film A maximum cooling of ΔT ~5 K was calculated. Liquid Film Thickness [µm] Using a constant piston crown wall temperature is justified. Δ Temperature [K] 0-5 Michael Heiss Sheet 12
Investigated Injection Timings high early injection low 20 CA late injection Michael Heiss Sheet 13
Formation of Liquid Film, late injection Liquid Film Thickness [µm] 1-Component Fuel Multicomponent Fuel Michael Heiss Sheet 14
Investigated Injection Timings high early injection low 20 CA late injection Michael Heiss Sheet 15
Formation of Liquid Film, early injection Liquid Film Thickness [µm] 1-Component Fuel Multicomponent Fuel Michael Heiss Sheet 16
Comparison with Smokemeter Measurements high The remaining wall film mass correlates with the measured smoke number low late injection early injection Michael Heiss Sheet 17
Multi-Component Liquid Film Composition, early injection injection high boiling temperature low-boiling temperature Components with lowest mass fractions in the fuel definition Michael Heiss Sheet 18
Vapour Sources, early injection 50 CA Michael Heiss Sheet 19
Mixture Preparation @ Spark Timing 717 CA, early injection 1-Component Fuel Multi-Component Fuel Lambda [-] 4% higher global lambda 12% higher lambda close to the spark plug Michael Heiss Sheet 20
Mixture Preparation @ Spark Timing 717 CA, late injection 1-Component Fuel Multi-Component Fuel Lambda [-] global lambda is equal lambda distribution is similar Michael Heiss Sheet 21
Conclusion The CFD simulation for an early injection timing showed significant differences between the single- and multi-component fuel approach. The intensified wall wetting and the remaining wall film mass after compression for early injection timings could only be represented correctly with the multi-component fuel definition. The fast wall film vaporisation of the single-component fuel led to an overestimation of the charge homogenisation and the fuel vapour fraction especially close to the spark plug. For operating points where only low wall film masses are to be expected the single-component approach is a reasonable simplification. Michael Heiss Sheet 22
Summary and Outlook At the test bench a sharp increase of soot emissions was measured when the start of injection exceeded a critical limit. With a 7 component fuel approach it was possible to calculate a remaining wall film mass after compression that correlated with the measured soot number. In contrast, the single-component fuel led to considerable differences in wall film mass and lambda distribution for an early injection. Therefore, the multi-component fuel approach has a big potential especially for operating points with intensified wall wetting e.g. cold start conditions and for a following combustion calculation where lambda needs to be as accurate as possible. For alternative fuels with a higher heat of vaporisation like ethanol, a moving mesh simulation with solid piston cells would be necessary to account for the intensified wall cooling. Michael Heiss Sheet 23
Thank you for your Attention! Michael Heiss michael.heiss@ifa.tuwien.ac.at