Scholarlink

<p>We examine the flow behavior around a transversely oscillating circular cylinder using various dimensionality reduction techniques. Specifically, Fourier analysis, Proper Orthogonal Decomposition (POD), Dynamic Mode Decomposition (DMD), and multi-resolution DMD (mrDMD) are employed. Numerical simulations are performed at a cylinder-diameter-based Reynolds number of 500 for a range of oscillation displacement amplitudes. The flow field exhibits well-documented wake patterns, such as 2S, 2P, and P+S, as well as intermittent transitions between these patterns at varying amplitudes. Dimensionality reduction becomes particularly effective when the force spectrum exhibits a dominant tonal character. Under these circumstances, the selection of the modal decomposition technique has minimal impact–all approaches yield comparable mode shapes for the dominant modes. However, when the flow undergoes intermittent pattern switching (e.g., between 2P and 2S), only mrDMD is able to <i>automatically</i> distinguish them as distinct modes. Nonetheless, if the temporal windows over which mode switching occurs are specified a priori, POD, DMD, and DFT are also successful.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Modal decomposition of flow behind a harmonically oscillating circular cylinder

  • 期刊论文

  • 发表日期：

    2025年12月8日

查看全文
  • Sudeep Menon,
  • Xingeng Wu,
  • Anupam Sharma

摘要

We examine the flow behavior around a transversely oscillating circular cylinder using various dimensionality reduction techniques. Specifically, Fourier analysis, Proper Orthogonal Decomposition (POD), Dynamic Mode Decomposition (DMD), and multi-resolution DMD (mrDMD) are employed. Numerical simulations are performed at a cylinder-diameter-based Reynolds number of 500 for a range of oscillation displacement amplitudes. The flow field exhibits well-documented wake patterns, such as 2S, 2P, and P+S, as well as intermittent transitions between these patterns at varying amplitudes. Dimensionality reduction becomes particularly effective when the force spectrum exhibits a dominant tonal character. Under these circumstances, the selection of the modal decomposition technique has minimal impact–all approaches yield comparable mode shapes for the dominant modes. However, when the flow undergoes intermittent pattern switching (e.g., between 2P and 2S), only mrDMD is able to automatically distinguish them as distinct modes. Nonetheless, if the temporal windows over which mode switching occurs are specified a priori, POD, DMD, and DFT are also successful.