Erscheint 6 Ausgaben pro Jahr
ISSN Druckformat: 1543-1649
ISSN Online: 1940-4352
Indexed in
AN OPTIMAL PREDICTION METHOD FOR UNDERRESOLVED TIME-MARCHING AND TIME-SPECTRAL
ABSTRAKT
The prevalence of multiscale phenomena in nonlinear dynamical systems often necessitates the use of reduced-order models. Similar in concept to the traditional view of spatial reduced-order models, under-resolved time-discretization schemes may be used to decrease the computational burden while sacrificing some level of accuracy. We present a framework based on the optimal prediction formalism that can be used to minimize the errors intrinsic to any timediscretization scheme. Models will be developed for the unresolved subgrid-scale dynamics and added to the governing equations. These subgrid-scale models will interact with the resolved timescales as the system evolves, rendering the optimal solution with respect to the chosen resolution. To demonstrate the technique, we study the van der Pol oscillator using a well-known time-marching scheme and a novel time-spectral method.
-
LaBryer Allen, Attar Peter J., Vedula Prakash, Optimal spatiotemporal reduced order modeling, Part I: proposed framework, Computational Mechanics, 52, 2, 2013. Crossref
-
LaBryer A., Attar P.J., Vedula P., Characterization of subgrid-scale dynamics for a nonlinear beam, Computers & Structures, 129, 2013. Crossref
-
LaBryer A., Attar P.J., Vedula P., Optimal spatiotemporal reduced order modeling of the viscous Burgers equation, Finite Elements in Analysis and Design, 79, 2014. Crossref
-
LaBryer A., Attar P. J., Vedula P., A framework for large eddy simulation of Burgers turbulence based upon spatial and temporal statistical information, Physics of Fluids, 27, 3, 2015. Crossref