Network-Theoretic Modeling of Fluid Flow

Abstract

A network-theoretic framework has been developed to model unsteady flows. Network analysis has been utilized to examine a wide variety of large-scale problems, including control of disease transmission, robust operation of power grids/internet, and uncovering brain functionality. The present investigation extends the network-theoretic approaches to describe the complex nonlinear behavior of unsteady fluid flows with particular focus on vortex interaction and modal energy transfer. Three canonical unsteady fluid flow problems are examined with network analysis: (1) discrete vortex dynamics, (2) modal interaction during wake formation in cylinder flow, and (3) two-dimensional isotropic homogeneous turbulence. Graph theory is utilized to extract important fluid flow interactions, which enabled the development of a sparsified vortex dynamics model. The network-based approach to examine the energy transfer between modal fluid flow structures has allowed for intuitive understanding of how perturbation or control input can alter the dynamics of the global fluid flow. At last, it was revealed that two-dimensional turbulence has an underlying scale-free network structure. The findings from the present study lay out a foundation to perform further network-based characterization and control to modify the collective behavior of unsteady fluid flows.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 29, 2015
Accession Number
AD1001358

Entities

People

  • Kunihiko Taira

Organizations

  • Florida State University

Tags

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Fluid Mechanics and Fluid Dynamics.
  • Neural Network Machine Learning.