Cycle-to-cycle variation quantification of flow fields using modified combined magnitude and relevance index

Cycle-to-cycle variations (CCV), defined by the inconsistency of combustion events across successive engine cycles despite identical nominal operating conditions, pose a significant challenge to stable operations, thereby limiting potential enhancements in engine efficiency and emissions reduction. Advancements in engine diagnostics have improved the ability to conduct detailed experimental studies on in-cylinder processes’ CCV. These investigations have identified key sources of CCV, such as the kinetic energy of the flow field under stoichiometric operating conditions, the interaction between flow and spray in direct-injection sparkignition (DISI) engines, and specific flow features in particular regions. Quantifying differences in vector fields is required to capture cyclic variations and confirm a successful mitigation in sources of CCV. However, the comparison of vector fields for quantifying cycle-to-cycle variation (CCV) remains a challenge. Vector field comparison, often conducted using metrics like the Relevance Index (RI) and Magnitude Similarity Index (MSI), allows for the quantification of differences between flow fields obtained from experimental measurements and numerical simulations. However, these metrics can yield misleading results in comparison of vector fields having relatively high varying velocity magnitude areas, such as areas near the vortex centers of in-cylinder turbulent flows. In addition, their high dependency on velocity magnitude can also result in a misleading focus on low-velocity magnitude areas. Previous works have introduced normalized metrics such as the Weighted Relevance Index (WRI) and the Weighted Magnitude Index (WMI), and their combination as the Combined Magnitude and Relevance Index (CMRI) with less sensitivity to such areas. However, CMRI, as a method for combining WRI and WMI, does not consider their significance equally, which can lead to errors in comparison of vector fields. In this paper, a comparison of advanced metrics used for comparing vector fields is ...