Controlled substructure identification for shear structures.

摘要

Shear structure models are widely used to model the dynamics of building structures; therefore, developing the techniques that can accurately identify the parameters of a shear structure plays a vital role in establishing efficient and reliable structural health monitoring systems for building structures.;In this dissertation, applying the "divide and conquer" strategy of substructure identification (SI), a series of innovative SI methods for shear structure are developed. A shear structure is divided into many two-story standard substructures. A novel inductive identification procedure is applied to identify the parameters of the whole structure from top to bottom. Numerical simulations verify that these substructure identification methods provide accurate identification results.;One of the most important features of these SI methods is that an approximate analytical expression for the identification error is obtained, which demonstrates that the identification accuracy is simply controlled by the frequency responses of the substructure near the substructure natural frequency. This important discovery provides the ability to easily improve the identification accuracy by appropriately changing the substructure responses via specially designed structural control systems. Several controlled substructure identification methods are proposed, using different structural control systems to improve the accuracy of the SI method. Furthermore, since the accuracy of the proposed controlled SI methods directly depends on the close-loop controlled structural responses rather than on the control systems themselves, these controlled SI methods are proven to be quite robust to possible control system errors.;To expand the applicability of the SI methods, a loop substructure identification method is proposed which makes use of the dynamic equilibrium of only one standard substructure to formulate a loop identification sequence and identify all parameters of that substructure once. Compared with the previous SI methods, the loop SI method is able to perform the structural identification of any part of a shear structure with only three floor acceleration responses; also importantly, the loop substructure identification can be carried out without knowing structural mass information.;Several shake table experiments are conducted on a two-story bench-scale test structure; the results show that the proposed SI methods can accurately identify the structural parameters and that, by using appropriately designed passive control system, the identification accuracy can be further improved.;Finally, a new approach is proposed to extend the SI methods originally developed for shear structures to more realistic frame structures. The study shows that the proposed approach is able to accurately structural damage of columns in a frame structure.;In summary, the SI methods developed in this dissertation are able to accurately identify the structural parameters of a shear structure, forming a solid foundation to design efficient SHM systems for building structures. Furthermore, combined with structural control systems, the proposed controlled SI methods not only further improve the accuracy of damage detection but also have a potential to enhance the performance of control systems to reduce the structural vibration by providing more accurate structural model for control algorithms design, both of which greatly enhance the safety and reliability of the structures.