Numerical Investigation of the Influence of different Valve Seat Geometries on the In-Cylinder Flow and Combustion in Spark Ignition Engines

Transcription

1 Institute for Combustion and Gas Dynamics Fluid Dynamics Numerical Investigation of the Influence of different Valve Seat Geometries on the In-Cylinder Flow and Combustion in Spark Ignition Engines Peter Janas 1, Usman Allauddin 2, Michael Pfitzner 2, Benjamin Böhm 3, Andreas Kempf 1 1 Universität Duisburg-Essen, Institute for combustion and gas dynamics, Germany 2 University Armed Forces Munich, Germany 3 Technical University Darmstadt, Germany

2 Engine Simulation Workflow Software Grid OpenFOAM-2.3.x Unstructured grids with local grid refinement Equidistant cells inside the cylinder Mesh motion Moving grids without topological changes Meshing Automated meshing + mapping Flow solver Compressible (pressure-based) Turbulence Smagorinsky (C s =0.062) Time Second-order implicit backward scheme discretization Numerical discretization Inlet/outlet boundary condition CDS-TVD for momentum TVD for convective scalar fluxes Time-varying pressure at the position of the pressure sensor Heat loss Isothermal walls (T intakeport =295 K, T walls =333 K) Fig.: Temperature during intake stroke ( x=0.25mm) Crevice Performance Included Cold flow cycle: 2-3 days; with combustion: 8 days (192 CPUs) 2

3 Optical Research Engine (TU Darmstadt) Cylinder head replaceable! Computational grids: ~every 5 CAD a new grid ( Mio cells) Mid valve plane mm Piston top-land crevice

4 Comparison of wall- and spray guided head Injector Injector Main Differences Intake Valve Diameter Spark-plug location Injector location Valve seat geometry Compression ratio 4

5 Experiments and boundary conditions Motored Case PIV 800 rpm Intake Pressure: ~1 bar Intake Temperature: 295 K IVO/IVC: -325 CAD / -125 CAD Max. valve lift: 9.5 mm Avg. of 2700 Cycles B A-A B Flow bench 3D data Magnetic Resonance Velocimetry Reynolds similarity exploited Fluid: Water Transparent 1:1 dummy engine Mimics -270 CAD (Intake) A B-B Ref.: Flow field measurement (Baum et al. Flow Turbulence Combustion, 2013) A Fired Case (Wall-guided only!) Fuel: Iso-octane (Φ=0.833) Spark Advance: 16 btdc 800 rpm Intake Pressure: ~1 bar Intake Temperature: 308 K 300 Cycles Ref.: 3D flow data around the valve Freudenhammer et al. SAE International Journal, 2015; Freudenhammer et al. Flow Turbulence and Combustion, 2014 Ref.: Carl-Philipp Ding, Laserbasierte Untersuchung der Flammenausbreitung in einem optisch zugänglichen direkteinspritzenden Ottomotor, Master Thesis,

6 Peak In-Cylinder Pressure (motored) Without crevice TDC Piston top-land crevice volume reduces the peak in-cylinder pressure by ~ 2bar! Similar pressure for the spray- and wall guided head

7 Ref.: Freudenhammer, Daniel, et al (2015): Intake Stroke Tumble symmetry plane Avg. of 15 cycles Avg. of 2700 cycles 7

8 Ref.: Freudenhammer, Daniel, et al (2015): Cold flow: LES vs. MRV - Intake Wall-Guided Spray-Guided Jet impingement Less recirculation *Avg. of 15 cycles *Avg. of 15 cycles 8

9 Ref.: Freudenhammer, Daniel, et al (2015): Compression Stroke: LES vs. PIV (motored) *Avg. of 15 cycles *Avg. of 2700 cycles 9

10 Tumble structure Gamma vortex identification planes Assumption:Sampling Rotational axis is perpendicular to the sampling plane!!!! Vortex identification algorithm 10

11 Auto-Correlation Functions: Wall and Spray smooth oscillating Ref.: Janas, Peter, et al. "On the Evolution of the Flow Field in a Spark Ignition Engine." Flow, Turbulence and Combustion (2016) 11

12 Auto-Correlation Functions: Wall and Spray Length scales Attachment with walls Change in the direction Free motion Ref.: Janas, Peter, et al. "On the Evolution of the Flow Field in a Spark Ignition Engine." Flow, Turbulence and Combustion (2016) 12

13 Combustion modelling Combustion model has to account for the ignition phase, flame kernel growth, free flame propagation, dilution of residual gases and wall effects. Flame propagation (c=0 unburnt and c=1 burnt): Flame Surface Density (*Keppeler et al.) : Ignition: Super-imposition of a spherical profile for progress variable Crevice effects: Reaction source term set to 0 in piston top-land crevice One-step chemistry Laminar flame speed: Metghalchi and Keck (Combustion and Flame, 1982) *Ref.: Keppeler et al. (Flow, turbulence and combustion 92.3, 2014) 13

14 Wall-Guided vs. Spray-guided (no residual gas) Flow In-cylinder field at Pressure -16 CAD Wall-guided head Averaged progress variable field -10 CAD Spray-guided head -5 CAD TDC Wall-guided head Avg. of 16 cycles Spray-guided head Avg. of 10 cycles Velocity profiles taken from cold flow analysis. Up to 10 CAD similar pressure due to the missing ignition model. Flame is burning slower towards the piston for the spray-guided head. Strong crevice effects after TDC. 14

15 Multi-Cycle simulation: Wall and Spray guided Wall-guided Spray-guided Spray-guided Wall-guided 15

16 Wall-Guided: Multi-Cycle simulation LES -5 TDC CAD Experiment *Avg. of 300 cycles *Avg. of 5 cycles Residual gases are decelerating the combustion progress (~15% IGR). In-cylinder pressure is in qualitative good agreement with the experiment up to the point where the flame reaches the liner. Cyclic variation caused by residual gas stratification and turbulence. 16

17 Mixing during Intake stroke Wall-guided head Spray-guided head 17

18 Institute for Combustion and Gas Dynamics Fluid Dynamics Piston top-land crevice volume increases wall heat transfer and changes the combustion phasing. Geometric variations in the valve seat lead to a complete different tumble formation. Length scales are decreasing with two distinct plateaus. Combustion is influenced by the residual gas stratification and flow field at ignition.

19 Institute for Combustion and Gas Dynamics Fluid Dynamics Thank you for your attention! Acknowledgements Ministry of Innovations, Science, and Research of the state Nord Rhine-Westphalia, Germany. Deutsche Forschungsgemeinschaft (DFG) Center for Computational Science and Simulation (CCSS) for computational resources