| Summary: | 碩士 === 朝陽科技大學 === 營建工程系 === 107 === During past few decades, structural health monitoring has attracted a lot attention. Structures may be subjected to seismic forces, wind loads and other effects over its lifetime use. Therefore, the structural parameters may be deviated from the design values due to the yielding or the fatigue of the material strength. In this regard, the dynamic characteristics may also be changed due to the damage of the structure. In order to realize the dynamic behavior of structural systems, we can determine the dynamic models and parameters by system identification techniques. System identification techniques also made possible to monitor the current state of the structure.
The traditional approach to earthquake resistant design is to design structures with sufficient strength capacity and the ability to deform in a ductile manner. Alternatively, newer concepts of structural control, including both passive and active control systems, have been growing in acceptance and may preclude the necessity of allowing for inelastic deformations in the structural system. A compromise between passive and active control systems has been developed recently in the form of semi-active control systems. Semi-active control systems maintain the reliability of passive control systems while taking advantage of the adjustable parameter characteristics of an active control system.
Yeh et al. proposed a control law called “Least Input Energy Method” (LIEM). The goal of LIEM is to minimize the input seismic energy to the superstructure by adjusting the isolation stiffness, so that the dynamic response of the structure can be mitigated. The proposed LIEM control law is applied to a semi-active isolation system called “Leverage-type Stiffness Controllable Isolation System” (LSCIS), and the isolation stiffness of the LSCIS can be controlled on-line by varying the pivot of its leverage [1].
In order to capture the behavior of LSCIS, this study builds up a numerical analysis program based on the numerical analysis method of LSCIS proposed by Yeh [1]. The validity of the proposed program is demonstrated by comparing the results of this study with the theoretical results of the LSCIS provided by Yeh [1].
A hybrid Genetic Algorithm developed by Wang [3] and Hong [4] is required to be applied to the identification of both the device and structural system. Accordingly, a hybrid Genetic Algorithm, merging the numerical analysis program associated with the LSCIS device has been developed in this study. By this hybrid computational strategy we can perform the parametric identification of the LSCIS device it self, and then implementing the identification techniques to the structure equipped with LSCIS.
The proposed algorithm has been applied to identify the system parameters of both the simulated LSCIS device and the simulated structure equipped with LSCIS device with or without noise contamination. Accordingly, the feasibility of the proposed new method has been verified. Finally, this identification algorithm has also been applied to the measured data from the experiments performed by Yeh [1, 2]. The comparisons has been made between the predicted response and the measured one for both the device and structural system equipped with those devices.
|
|---|
