UBC Theses and Dissertations
Experimental and numerical simulation of charge motion in internal combustion engines Anetor, Lucky
A combined study based on Laser Doppler velocimetry measurements and numerical simulation was used to investigate the effects of piston-bowl geometry and intake swirl level on the structure and evolution of the flow field near the top dead centre position of an internal combustion engine. The University of British Columbia Rapid Intake and Compression Machine (UBCRICM), was used for the experimental studies, while numerical calculations were performed with the KIVA-Il code. The KIVA-Il code is an incylinder fluid dynamics flow computer program, incorporating the following submodels; spray dynamics, species transport, mixing, chemical reactions and heat release rate. The effects of turbulence are represented by either the k-e or the subgrid scale (SGS) turbulence models. The k-e turbulence model option was used in the present study. The results of this study show that high-shear regions and the consequent turbulence production occurred near the bowl entrance around top dead centre of compression. These regions were created before top dead centre in both chambers by the interaction of swirl and squish. The flow parameters (mean radial and tangential velocities and turbulence intensity) resulting from the re-entrant bowl configuration were found to be consistently higher than those generated by the bowl-in-piston arrangement. The com putational results were compared to the experimental data. A very good agreement was obtained between predicted and measured mean radial and tangential velocity distribu tions. The numerical code was found t under estimate the measured turbulence intensity values in both chamber designs, however, a reasonable level of agreement was obtained with both chamber configurations. This study also highlighted the importance of accurately specifying the initial and boundary conditions. It was found that the very impressive level of agreement between measured and computed results was largely due to the fact that the initial and boundary conditions were accurately modelled to mimic actual experimental conditions. The investigation of the evolution of certain flow and turbulence quantities which could not be obtained experimentally was also undertaken with the KIVA-Il code. A comprehensive discussion of these flow quantities was presented and the various ways they are likely to affect engine performance in terms of combustion were highlighted.