LES Multi-cycle Analysis of a High Performance GDI Engine

The paper reports the application of LES multi-cycle analysis for the characterization of cycle to cycle variability (hereafter CCV) of a highly downsized DISI engine for sport car applications. The analysis covers several subsequent engine cycles operating the engine at full load, peak power engine speed. Despite the chosen engine operation is usually considered relatively stable, relevant fluctuations were experimentally measured in terms of in-cylinder pressure evolution and combustion phasing. On one hand, despite the complex architecture of the V-8 engine, the origin of such CCV is considered to be poorly related to cyclic fluctuations of the gas-dynamics within the intake and exhaust pipes, since acquisitions of the instantaneous pressure traces at both the intake port entrance and exhaust port junction by fast-response pressure measurements over 250 subsequent engine cycles showed almost negligible differences in both amplitude and phasing compared to those within the cylinder. On the other hand, being the combustion affected by a complex chain of preceding factors (air admission during the intake stroke, variations in the residual gas fraction, generation of complex turbulent flow structures, fuel injection and dispersion in the combustion chamber and subsequent mixing, interaction between the spark discharge and the surrounding local flow pattern, etc.) a clear understanding of the actual origin of cyclic variability is far from being trivial. LES CFD simulations can therefore become a very powerful tool to help investigating the possible causes of such cyclic variations, since detailed analyses of both global and local parameters can be carried out on an almost unlimited set of available virtual measurements. In the first part of the paper, the modeling framework is presented and considerations on the adopted numerical strategy are presented, with particular emphasis on grid size, grid distribution and numerical parameters. Subsequently, LES results are analyzed and discussed in order to understand the cycle-to-cycle variations through the use of correlation coefficients between global/local flow variables in order to highlight the major causes of CCV and establish a possible hierarchy among the analyzed quantities. Finally, criticalities of the currently adopted approach and possible enhancements are briefly discussed at the end of the paper. The results presented in the paper clearly highlight the potential of the modeling methodology to help understanding the origin of CCV as well as to address possible engine optimizations to limit the cyclic dispersion.