Modal Decomposition of Wake Flow Behind a Circular Cylinder

Abstract

This study demonstrates the effectiveness of Proper Orthogonal Decomposition (POD) as a method for analyzing flow structures and as a data reduction tool for both numerical simulations and experimental techniques. The reduced data can be integrated to data driven inference of partial differential equations which can lead to reduced order modeling. By focusing on the flow around a cylinder, we show that POD efficiently identifies the most energetic modes, allowing for the flow's reconstruction with a limited dataset. Our findings indicate that the first 10 POD modes appear in pairs with similar spatial patterns, frequency spectra, and energy contributions. These modes encapsulate nearly 92% of the total energy, highlighting the method's precision in capturing essential dynamics. Specifically, modes 1 and 2 correspond to streamwise vortex shedding, while modes 3 and 4 relate to the lateral flow motion. The reconstructed flow, based on these dominant modes, closely matches numerical analysis results, validating POD's utility in developing reduced order models and enhancing our understanding of fluid-structure interactions.