Challenges in variable-resolution modeling of turbulence: Fluctuating flow-field physics and characterization
Conference Paper
Overview
Identity
View All
Overview
abstract
While variable-resolution (VR) methods are emerging as popular turbulence computation tools, many challenges remain. One important challenge is to ensure the physical fidelity of the VR fluctuations. The primary purpose of variable-resolution turbulence simulations is to resolve important scales of motion but the physical attributes of the fluctuations are rarely examined. Based on the paradigm that variableresolution turbulence computation is tantamount to Direct Numerical Simulation (DNS) of variable-viscosity turbulence, we propose a framework for characterizing and validating VR spatial and temporal fluctuations. Exploiting the DNS analogy, we suggest that the VR fluctuations must scale according to Kolmogorov refined hypotheses and local dissipation-scales similarity arguments. We perform partially-averaged Navier-Stokes (PANS) VR computations of flows past a circular cylinder and backward-facing step and demonstrate that: (i) the normalized PANS local dissipation length and time scale distributions are indeed nearly invariant at different computed resolutions and (ii) the invariant VR-distribution form is somewhat similar to that in decaying isotropic turbulence. The results indicate that the classical scaling laws are indeed meaningful for VR computations and can be used to establish the physical fidelity of VR fluctuations. Copyright 2011 by Sharath Girimaji.
