Intrinsically Efficient and Accurate Viscous Simulations via Hyperbolic Navier-Stokes Systems
Abstract
Unstructured-grid methods are essential for computations with complex geometries such as rotorcraft simulations, but its potential has been limited by a higher cost than structured-grid methods as well as inaccuracy in gradient predictions (e.g., diffused or oscillatory vorticity predictions). Grid irregularities are hard to avoid particularly once grid adaptation is performed, which is a critical technique especially for high-order unstructured-grids methods to be practical. Current state-of-the-art Navier-Stokes (NS) codes are known to produce highly erratic viscous stress and heating distributions. Resolution of these problems is very important for justifying the use of high-fidelity models in aerodynamic design and optimization. In this project, we address these issues by developing a Navier-Stokes solver based on a novel first order hyperbolic system method. The new solver is expected to yield O(1/h) acceleration in convergence over existing solvers, where h is the typical grid spacing, as well as achieve high-accuracy in auxiliary quantities, viscous stresses, heating rates, and vorticity, on unstructured grids. These improvements will be achieved by the new method in which the Navier-Stokes equations are discretized as a first-order hyperbolic system including the auxiliary quantities as additional variables. The new code will enable complex large-scale simulations with the current hardware and meet the challenge of highly efficient and accurate multi-scale unsteady aerodynamic computations of Army's interest: vortex-dominated flows, separated flows, wake interaction, and dynamic stall of rotor-craft, helicopter blades, high-speed missiles, gun-launched projectiles, micro air vehicles, and micro adaptive flow control, which require especially accurate vorticity predictions. The new solver is implemented in the framework of a practical flow solver used by Army, NASA's fully unstructured and parallel 3D RANS code, FUN3D.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 24, 2016
- Accession Number
- AD1058159
Entities
People
- Hiroaki Nishikawa
Organizations
- National Institute of Aerospace