A Comparative Assessment of Variant k-epsilon Models for Engine Flow Applications

In this research, the performance of non-linear k-epsilon turbulence model in resolving the time delay between mean flow changes and its proportionate turbulent dissipation rate adjustment was investigated. For this purpose, the ability of Launder-Spalding linear, Sug a non-linear, Yak hot RNG, Rietz modified RNG k-epsilon models were compared in estimation of axial mean velocity profile and turbulent integral length scale evolution during engine compression stroke. Computed results showed that all the models can predict the acceptable results for velocity profile but for turbulent integral length scale curve Suga non-linear model is in a good agreement with Rietz modified RNG model prediction, while other models showed a different unrealistic behavior. It is concluded that, although in some features all the models have proper results, for integral length scale just modified RNG and non-linear models depict correspondence with experimental reported data.