Large eddy simulations of the influence of piston position on the swirling flow in a model two-stroke diesel engine

Abstract

Received (Day Month Year) Revised (Day Month Year)Purpose-the purpose of this paper is to study the effect of piston position on the in-cylinder swirling flow in a simplified model of a large two-stroke marine diesel engine.Design/Methodology/Approach-Large Eddy Simulations with four different models for the turbulent flow are used: a one-equation model, a dynamic one-equation model, a localized dynamic one-equation model and a mixed-scale model. Simulations are carried out for two different geometries corresponding to 100 % and 50 % open scavenge ports.Findings-It is found that the mean tangential profile inside the cylinder changes qualitatively with port closure from a Lamb-Oseen vortex profile to a solid body rotation while the axial velocity changes from a wake-like profile to a jet-like profile. The numerical results are compared with particle image velocimetry measurements (?) and in general we find a good agreement.Limitations/implications-Considering the complexity of the real engine, we designed the engine model using the simplest configuration possible. The setup contains no moving parts, the combustion is neglected and the exhaust valve is discarded.Originality/value-Studying the flow in a simplified engine model, the setup allows studies of fundamental aspects of swirling flow in a uniform scavenged engine. Comparing the four turbulence models, the local dynamic one-equation model is found to give the best agreement with the experimental results.

Turbulent swirling flow in a dynamic model of a uniflow-scavenged two-stroke engine

“…The swirling in-cylinder flow in uniflow-scavenged engines has been the subject of earlier works. Investigations have been based on both CFD (Sung and Patterson 1982;Diwakar 1987;Uzkan 1988;Goldsborough and Blarigan 2003;Obeidat et al 2014;Sigurdsson et al 2014) and experiments (Percival 1955;Nakagawa et al 1990;Sher et al 1991;Haider et al 2013;Ingvorsen et al 2012Ingvorsen et al , 2013. Due to the costs and limitations of experimental investigations on full-scale engines, the majority of the experimental work have been performed on simplified scale models, and often under steady-flow conditions.…”

An implicit boundary approach for viscous compressible high Reynolds flows using a hybrid remeshed particle hydrodynamics method

“…al. [62] boundary condition formula. To be consistent with the 26 • flow angle measured in [63], a uniform radial and tangential 275 velocity at the inlet is defined with a constant radial speed V r,i = 0.23V b and a constant tangential speed V θ,i = 0.11V b .…”

“…At the outlet we require a zero velocity gradient. We simulate the flow using the hrSPH method on a grid with approximately 4 million cells, whereas 8 million cells were used in Obeidat [62] to achieve satisfactory results compared to the experimental ones. The time 280 step is chosen such that the solution is stable and the Courant number remains below 1 (u∆t/∆x < 1).…”

“…This setup nevertheless allows us to study fundamental aspects of a swirling flow in a uniform scavenged engine. The results are compared with those of another study by Obeidat[62].…”

“…al. [62], additionally, we are using the simplest turbulence model (Smagorinsky). When enough resources are put down to develop a Large eddy simulation model, we expect that the hrSPH method will be able to give a more accurate results for such a high Reynolds number and comparatively coarser mesh.…”