Reduced Order Models in Analysis of Stochastically Parametered Linear Dynamical Systems

Abstract This study focusses on the development of reduced order models, which minimize the computational costs without compromising on the accuracy in the numerical analysis of stochastically parametered linear dynamical systems. A scheme based on polynomial chaos expansion (PCE) and system equivalent reduction expansion process (SEREP) has been developed that enable formulation of reduced order models. Further measures for enhancing the computational efficiency include using sparse grids in conjunction with code parallelization. Interfacing algorithms have been developed that enable finite element (FE) modeling of complex systems using commercial FE softwares and the developed codes.