On the influence of the small-scale flow on cycle-to-cycle variations in charge diluted spark-ignition engines

Abstract

Charge dilution via exhaust gas recirculation (EGR) enhances efficiency and reduces nitrogen oxide emissions in spark-ignition engines but intensifies cycle-to-cycle variation (CCV), compromising stability. This study experimentally assesses flow-induced CCV drivers under charge dilution using homogeneous mixtures and artificially prepared EGR in an optically accessible spark-ignition engine, eliminating inhomogeneity effects. Residual gas effects are excluded by using a skip-firing scheme. High-speed particle image velocimetry (PIV) and flame imaging are used to capture the flow evolution and flame propagation. This work examines turbulence–flame interactions, identifying distinct flow conditions for fast and slow flame development during compression and ignition, with and without EGR. The flame size is analyzed in relation to cycle outcomes, while global and local flow properties – including kinetic energy, turbulent kinetic energy, and strain – are systematically evaluated. In mixtures with EGR, small differences in the kinetic and turbulent kinetic energy are found to be sufficient to cause differences in the flame propagation. Principal component analysis of the strain tensor highlights different behaviors for cycles of fast and slow flame propagation. Local flow conditions at the spark plug emerge as a decisive factor in both cases, with the orientation of the velocity vector at the spark gap and the turbulent kinetic energy exhibiting a notable influence on CCV.

Novelty and Significance

This work investigates cycle-to-cycle variations (CCV) in an optical spark-ignition engine with exhaust gas recirculation (EGR). A key novelty of this study is the systematic analysis and comparison of bulk and fluctuating flow structures with and without EGR, achieved by using artificially prepared exhaust gas and skip-firing to eliminate mixture-induced CCV and isolate the effects of fluctuating turbulent flow structures. While previous work has focused on the bulk flow neglecting the fluctuating components, this work employs conditioned statistics and empirical mode decomposition are applied to separate coherent structures and turbulent fluctuations from the bulk flow. A significant finding is that the importance of the local and global distribution of these structures changes depending on the presence of EGR.