The proper orthogonal decomposition (POD) method is applied to analyze the particle image velocimetry (PIV) measurement data and large eddy simulation (LES) result from an in-cylinder turbulence flow field in a four-valve direct injection spark ignition (DISI) engine. The instantaneous flow fields are decomposed into four parts, namely, mean field, coherent field, transition field and turbulent field, respectively, by the POD quadruple decomposition. The filtering method for separating the four flow parts is based on examining the relevance and correlations between different flow fields reconstructed with various POD mode numbers, and the corresponding reconstructed fields have been verified by their statistical properties. Then, the in-cylinder flow evolution and cycle-to-cycle variations (CCV) are studied separately upon the four field parts. Results indicate that each one of the four field parts exhibits its own flow characteristics and has close connection with others. Furthermore, the mean part contains the most kinetic energy of the entire flow field and represents the bulk flow of the original in-cylinder velocity field; the CCV in this part could almost be neglected, while the coherent field part contains larger scale structures and the most fluctuating energy, and possesses the highest CCV level among the four parts.
Rapid Reconstruction of Simulated and Experimental Temperature Fields Based on Proper Orthogonal Decomposition
