Investigation of boundary condition and field distribution effects on the cycle-to-cycle variability of a turbocharged GDI engine using LES

The paper reports some preliminary results of a numerical activity aiming at characterizing cycle-to-cycle variability of a highly-downsized turbocharged DISI engine for high-performance car applications, using a customized version of the commercial software Star-CD licensed by CD-adapco. During experimental investigations at the engine testbed, high cycle to cycle dispersion was detected even for relatively stable peak-power / full-load operations of the engine, thus limiting the overall engine performance. Despite the complex architecture of the V-8 engine, the origin of such cyclic variability could not be related to cyclic fluctuations of the gas-dynamics within the intake and exhaust pipes. Several subsequent acquisitions of the instantaneous pressure traces were measured at both the intake port entrance and exhaust port junction, showing almost negligible differences in terms of both amplitude and phasing compared to those within the cylinder. In order to explore the potentials of the LES application to the analysis and understanding of the cycle-to-cycle variability, which notoriously strongly limits the overall engine performance, a numerical activity is carried out using full-cycle LES simulations over several subsequent engine cycles. Despite the very early stage of the investigation, two main issues are addressed in the paper: the analysis of the possible causes originating the high cycle to cycle variability and the influence of the boundary conditions on the predicted cyclic dispersion. Concerning the former aspect, a detailed investigation of local and global instantaneous fields is carried out aiming at identifying both a possible hierarchy of responsibilities on one side and limitations and possible improvements of the adopted numerical procedure on the other side. Concerning the latter aspect, a set of simulations, performed applying cycle-independent averaged experimental conditions, is compared to those resulting from the application of cycle-specific variable pressure traces at both intake and exhaust sides, in order to analyze the influence of port fluctuations on the in-cylinder pressure history. Results are also qualitatively compared to those resulting from a multi-cycle RANS simulation using the same grid-size, in order to better highlight the superior LES potentials.