Reduced Order Modeling of Unsteady Flows Using Sparse Bases

Abstract

This effort will leverage sparse coding principles to construct accurate, robust, and efficient reduced order representations of the Navier-Stokes equations for sea-based aviation problems. Specifically, the effort will seek to identify a sparse set of modes in unsteady, nonlinear flow fields. The sparse basis approach will be applied to a variety of datasets ranging from those computed with Direct Numerical Simulations (DNS) and Large Eddy Simulations (LES) on structured meshes to Detached Eddy Simulation (DES) and Unsteady Reynolds Averaged Navier-Stokes (URANS) on unstructured meshes. Initial applications will include flow predictions for canonical configurations - oscillating cylinders, airfoils, backward facing steps, and combinations of these. Using these flow fields, the PI will seek both improved understanding on the unique properties of sparse bases in the context of reduced order solution of the Navier-Stokes equations, and adaptive procedures for non-stationary mean flow conditions.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 12, 2016
Source ID
N000141612620

Entities

People

  • Jack J. McNamara

Organizations

  • Office of Naval Research
  • Ohio State University
  • United States Navy

Tags

Fields of Study

  • Physics

Readers

  • Computational Fluid Dynamics (CFD)
  • Neural Network Machine Learning.