| In vibration-based model updating, the finite element model is iteratively modified to ensure its vibration properties reproduce the measured counterparts in an optimal way. Model updating of large-scale structures is usually expensive in terms of computation time and memory, since the finite element model of a large-scale structure generally consists of a large number of degrees of freedom and many uncertain parameters. This PhD study develops a forward and an inverse substructuring approaches for model updating of large-scale structures.;In the forward substructuring approach, the eigenproperties of the partitioned substructures are assembled to recover the eigensolutions and eigensensitivities of the global structure, which are tuned to reproduce the experimental measurements through an optimization process. Kron's substructuring method is improved by modal truncation technique for eigensolutions. This improvement not only reduces the computational endeavor in extracting the eigenmodes for the substructures, but also produces a much smaller eigenequation. The improved substructuring method for eigensolutions is subsequently extended to calculate the first-order and high-order eigensensitivities with respect to elemental parameters. The eigensensitivity of the global structure are determined from the derivative matrices of only particular substructures that contain the designed elements, thus realizing a significant reduction in computational cost. In consequence, the substructure-based eigensolutions and eigensensitivities are successfully applied in the practical model updating process. As the accurate eigensolutions and eigensensitivities are sometimes required, such as at the final steps of the model updating procedure, an iterative scheme is proposed to calculate the accurate eigensolutions and eigensensitivities using only a few master modes.;Afterwards, an inverse substructuring approach is developed to extract substructural flexibility from the experimental modal data. As a result, the focused substructure is treated as an independent structure to be updated directly using a conventional model updating method, thus accelerating the optimization process significantly. This inverse substructuring approach allows for the focused substructures to be updated directly based purely on the measurements taken in the local area.;The proposed substructuring-based model updating approaches have been successfully applied to a few numerical, laboratory, and real structures. |
