Coupling time-resolved PIV flow-fields and phase-invariant proper orthogonal decomposition for the description of the parameters space in a transparent Diesel engine

Abstract

Particle image velocimetry (PIV) results obtained from one transparent Diesel engine with a cylindrical bowl-in piston are investigated in this paper. Beyond the standard statistical description of in-cylinder flows, time-resolved PIV (TRPIV) now allows to access the in-cycle evolution. However, due to experimental and data storage limitations, the number of cycles obtained with TRPIV is evidently too low to perform any reliable statistical analysis and this is a general problem when TRPIV is used. However, this in-cycle information is very useful and should be associated in an optimal way to classical statistical data bases obtained in such engines. To overcome this issue, a coupling approach between TRPIV data and phase-invariant proper orthogonal decomposition (POD) applied on statistically converged data is proposed. To describe the large-scale motions and to reveal the low order flow dynamics, it is necessary to combine in an optimal way the statistically converged PIV datasets and the limited number of in-cycle TRPIV measurements. The POD modes resulting from the decomposition of the first dataset are used as an energetic basis for the projection of the time-resolved flow-fields during the whole compression stroke. The in-cycle evolution of the POD random coefficients is analysed. This method gives a low order description of the evolution of the compressed flow structure and dynamics and we prove that the fluctuations associated with both the intensity and the displacement of the swirling motion increase during the compression stroke, particularly during the transfer in the bowl. Normalized projections associated with the fluctuations identified are finally introduced to describe the parameters space for the available number of flow realisations. We show that they vary over the full parameter space during one compression.