KY EPSCoR: Data-Driven Adaptive Reynolds-Averaged Navier-Stokes K--Omega Models for Unsteady Turbulent Flow

Grants and Contracts Details


Turbulent flow arises in a vast array of NASA technologies ranging from flow around launch and entry vehicles to internal flow within propulsion and life-support systems. Thus, computational techniques for accurate turbulent-flow simulation can advance numerous NASA missions. Direct numerical simulation (DNS) is effective for turbulent flow; however, the precision of DNS is unavailable for complex geometries or high Reynolds numbers due to its immense computational cost. Instead, modeling approaches such as large-eddy simulation (LES), detached-eddy simulation (DES), and Reynoldsaveraged Navier-Stokes (RANS) must be used for these flows. The optimal parameters of these turbulence models are generally problem dependent and often determined using trial-and-error calibration with experimental or higher-fidelity numerical results. The objective of this project is to develop a data-driven adaptive CFD model for simulating unsteady turbulent flow, where partial-but-incomplete flow-field data is available. This project leverages our recently developed data-driven adaptive RANS k—ù (DARK) model for steady turbulent flow simulation. Thus, we aim to advance DARK by extending it to handle unsteady turbulent flow, and improving its computational efficiency. A key enabling technology for these advances is the PI's recently developed retrospective cost adaptation (RCA) approach. We will develop an improved DARK model by implementing RCA for DARK-model adaptation. We will test this improved model on several canonical test cases for unsteady turbulent flow.
Effective start/end date6/1/17 → 5/31/18


  • KY Council on Postsecondary Education


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.