A novel substructure coupling technique based on the proper orthogonal decomposition method is presented for the midfrequency range vibration of linear dynamical systems with parameter uncertainty. For a given frequency band, the methodology permits the derivation of an adaptive basis for each subsystem and the construction of a reduced-order model of the global structure. The formulation is directed toward the efficient probabilistic characterization of model-based predictions in the framework of a stochastic finite element method. The efficiency of the substructure method has been contrasted both from the viewpoint of adopting free-free and fixed-fixed substructure proper orthogonal modes in order to arrive at a reduced subsystem model. The distinction as well as similarity of the present methodology with the component mode synthesis is also pointed out: The proper orthogonal modes are obtained from both frequency- and time-domain approaches, and their suitability is discussed in relation to the behavior of a specific system. The substructure approach elegantly integrates with a version of stochastic finite elements based on orthogonal decompositions and projections of stochastic processes.

Additional Metadata
Persistent URL dx.doi.org/10.1121/1.1558374
Journal Journal of the Acoustical Society of America
Sarkar, A, & Ghanem, R. (Roger). (2003). A substructure approach for the midfrequency vibration of stochastic systems. Journal of the Acoustical Society of America, 113(4 I), 1922–1934. doi:10.1121/1.1558374