It is now well established that highly turbulent flows in engines are responsible for cycle-to-cycle aerodynamic variations particularly in cylinder. These dispersions may lead to macroscopic changes in flow pattern from one cycle to another one, which are quite different from “true turbulence” (in fact small scale turbulent fluctuations) used to optimize combustion.
It is of a great importance to decrease these variations which may effect reliability of combustion concept and particularly when looking to decrease idle engine speed. The difficulty in separating “true turbulence” from cyclic variations is that there is no limit between two kinds of fluctuations.
To analyze aerodynamic, PIV is well adapted to study these phenomena and is now currently used in development process, allowing to measure instantaneous aerodynamics field to determine for instance vortex center position for each cycle. From these data, to be able to separate “true turbulence” from cyclic variations, different advanced mathematical techniques have been tested since several years.
Work performed here is based on Proper Orthogonal Decomposition (POD). Main results show that using this analysis allow to provide a spatial representation of “true turbulence” and cycle-to-cycle dispersions fields. Future developments are finally discussed.