Item

Abstract

Streaming Dynamic Mode Decomposition (sDMD) is a low-storage versionof dynamic mode decomposition (DMD), a data-driven method to extractspatiotemporal flow patterns. Streaming DMD avoids storing the entire datasequence in memory by approximating the dynamic modes through incremen-tal updates with new available data. In this paper, we use sDMD to identify andextract dominant spatiotemporal structures of different turbulent flows, requir-ing the analysis of large datasets. First, the efficiency and accuracy of sDMDare compared to the classical DMD, using a publicly available test dataset thatconsists of velocity field snapshots obtained by direct numerical simulation of awake flow behind a cylinder. Streaming DMD not only reliably reproduces themost important dynamical features of the flow; our calculations also highlightitsadvantageintermsoftherequiredcomputationalresources.Wesubsequentlyuse sDMD to analyse three different turbulent flows that all show some degreeof large-scale coherence: rapidly rotating Rayleigh–Bénard convection, horizon-tal convection and the asymptotic suction boundary layer (ASBL). Structures ofdifferent frequencies and spatial extent can be clearly separated, and the promi-nent features of the dynamics are captured with just a few dynamic modes. Insummary, we demonstrate that sDMD is a powerful tool for the identification ofspatiotemporal structures in a wide range of turbulent flows.