Optimal reduced-order modeling for nonlinear distributed parameter systems

Author

Borggaard, Jeff

Author_Institution

Dept. of Math., Virginia Tech, Blacksburg, VA

fYear

2006

fDate

14-16 June 2006

Abstract

A method to develop reduced-order models for nonlinear distributed parameter systems is studied. The method is based on Galerkin projection, but the reduced-basis vectors are optimal for the dynamic model, found by minimizing the error between given full-order simulation data and the reduced-order model. This is achieved by formulating the basis selection problem as an optimal control problem with the reduced-order model as a constraint. This methodology allows a natural extension of reduced-order modeling ideas to nonlinear systems. A numerical experiment comparing the optimal reduced-order model to the popular proper orthogonal decomposition method is provided

Keywords

Galerkin method; Karhunen-Loeve transforms; distributed parameter systems; nonlinear control systems; nonlinear dynamical systems; optimal control; reduced order systems; Galerkin projection; basis selection; dynamic model; nonlinear distributed parameter systems; optimal control problem; optimal reduced-order modeling; orthogonal decomposition; reduced-basis vectors; Contracts; Differential equations; Distributed parameter systems; Hydrogen; Large-scale systems; Mathematics; Nonlinear systems; Reduced order systems; Vectors; Weather forecasting;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, 2006

Conference_Location

Minneapolis, MN

Print_ISBN

1-4244-0209-3

Electronic_ISBN

1-4244-0209-3

Type

conf

DOI

10.1109/ACC.2006.1656372

Filename

1656372