A predictive wall model for large-eddy simulation based on optimal control techniques

Abstract

Wall models for large-eddy simulation (LES) based on optimal control theory have so far been nonpredictive due to the need to prescribe a known mean velocity profile to the controller. In this study, LES is coupled with a near-wall Reynolds-averaged Navier–Stokes (RANS) model that provides a target velocity for the cost function. For the wall model to be accurate and robust, the LES and RANS must not only be tied together via the controller but directly coupled to each other through boundary conditions. The method proves to be accurate and robust over a wide range of Reynolds numbers in a plane channel flow. It is shown that the control reacts only locally in all spatial directions, justifying the current control formulation and suggesting directions for future model development. Further, instantaneous velocity fields of the coarse LES indicate that the dynamics of the near-wall flow are very dependent on the computational grid, demonstrating that a control strategy is required in addition to physical reasoning for wall modeling.

Document Details

Document Type
Pub Defense Publication
Publication Date
Jun 01, 2008
Source ID
10.1063/1.2930673

Entities

People

  • Jeremy A. Templeton
  • Meng Wang
  • Parviz Moin

Organizations

  • Air Force Office of Scientific Research
  • Stanford University

Tags

Fields of Study

  • Physics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Computational Fluid Dynamics (CFD)
  • Systems Analysis and Design