Decomposition of wake dynamics in fluid-structure interaction via low-dimensional models

Abstract

We present a dynamic decomposition analysis of the wake flow in fluid-structure interaction (FSI) systems under both laminar and turbulent flow conditions. Of particular interest is to provide the significance of low-dimensional wake flow features and their interaction dynamics to sustain the free vibration of a square cylinder at a relatively low mass ratio. To obtain the high-dimensional data, we employ a body-conforming variational fluid-structure interaction solver based on the recently developed partitioned iterative scheme and the dynamic subgrid-scale turbulence model for a moderate Reynolds number. The snapshot data from high-dimensional FSI simulations are projected to a low-dimensional subspace using the proper orthogonal decomposition (POD). We utilize each corresponding POD mode for detecting features: the vortex street, the shear layer and the near-wake bubble. We find that the vortex shedding modes contribute solely to the lift force, while the near-wake and shear layer modes play a dominating role to the drag force. We further examine the fundamental mechanism of this dynamical behavior and propose a force decomposition technique via low-dimensional approximation. We ascertain quantitatively that the shear layer feeds the vorticity flux to the wake vortices and the near-wake bubble during the wake-body synchronization. Based on the decomposition of wake dynamics, we suggest an interaction cycle for the frequency lock-in during the wake-body interaction, which provides the inter-relationship between the high amplitude motion and the dominating wake features. Through our investigation of wake-body synchronization below critical Re range, we discover that the bluff body can undergo a synchronized high-amplitude vibration due to flexibility-induced unsteadiness. The interaction cycle for the wake-synchronization is found to be valid for the turbulent wake flow.

arXiv comment42 pages, 20 figures

References in corpus (1)

• An Efficient Deep Learning Technique for the Navier-Stokes Equations: Application to Unsteady Wake Flow Dynamics

Cited by in corpus (5)

• Assessment of unsteady flow predictions using hybrid deep learning based reduced order models
• Three-dimensional deep learning-based reduced order model for unsteady flow dynamics with variable Reynolds number
• A Parametric and Feasibility Study for Data Sampling of the Dynamic Mode Decomposition: Spectral Insights and Further Explorations
• The Linear-Time-Invariance Notion of the Koopman Analysis-Part 2: Physical Interpretations of Invariant Koopman Modes and Phenomenological Revelations
• Aeroelastic mode decomposition framework and mode selection mechanism in fluid-membrane interaction