ARTICLE:

• REDUCED-ORDER MODELING OF TURBULENT FLOWS IN MULTISCALE DOMAINS  download article

J. Rambo
Georgia Institute of Technology, Atlanta, USA

Y. Joshi
Georgia Institute of Technology, Atlanta, USA

ABSTRACT

Reduced-order models of stationary turbulent flows capable of accurate flow field predictions are developed using the proper orthogonal decomposition and the methodology is extended to approximate the pressure field. This procedure can be used to efficiently span length scales by creating component level reduced-order models and assembling them together through a flux matching procedure to construct computationally efficient models of the detailed transport processes of large-scale systems for further analysis. Reduced-order models are coupled with lumped parameter models to create a modular system representative of an air-cooled electronics cabinet. The results indicate accurate approximations of turbulent flows across various domains can be obtained using flow network modeling procedures to assemble component models and efficiently bridge length scales.

 download article

TRB-10 pages

« Previous article         Next article »