Piston Bowl Geometry Effects on Combustion Development in a High-Speed Light-Duty Diesel Engine 2019-24-0167
In this work, we studied the effects of piston bowl design on combustion in a light-duty direct-injection diesel engine. Two bowl designs were compared: a conventional, omega-shaped bowl and a stepped-lip piston bowl. Experiments were carried out in the SNL single-cylinder optical engine facility, with a medium-load, mild-boosted operating condition featuring a pilot+main injection strategy. CFD simulations carried out with the FRESCO platform featuring full-geometric modeling of the engine, were validated against measured in-cylinder performance as well as soot natural luminosity images. Differences in combustion development were studied using the simulation results, and sensitivities to in-cylinder flow field (swirl ratio) and injection rate parameters were also analyzed. In-cylinder mixture formation analysis showed that ignition of the pilot injection mixture develops nearly as a homogeneous adiabatic reactor, being mostly advected, not mixed, by the bowl’s swirling motion, while its timing is influenced by the local flow field. Local in-cylinder flow is also more crucial than injection parameters to igniting the main injection’s premixed fuel, as it determines the relative overlap with the high-temperature pilot ignited mixture. Bowl geometry effects drive diffusive and late-cycle combustion, as structural differences of the main injection spray flames appear since the different impact geometries at the piston bowl rim. Better air utilization with the stepped-lip geometry, thanks greater azimuthal spreading at the rim, a strong recirculating vortex in the squish, and better mixing/lower swirl in the bowl, is responsible for better late-cycle combustion efficiency and lower soot emissions. A comparison between full- and sector-mesh CFD approaches also showed that flame structure is geometry-driven and well-captured by the sector, but combustion predictions are deteriorated by the sector symmetry’s inaccurate local flow field representation.
Federico Perini, Stephen Busch, Kan Zha, Rolf Reitz, Eric Kurtz
University of Wisconsin-Madison, Sandia National Laboratories, Ford Motor Company
14th International Conference on Engines & Vehicles