| Summary: | 碩士 === 國立交通大學 === 土木工程系所 === 108 === Earthquake disaster is one of the major threat to nationals’ lives and properties. In recent years, researches on structural control combining earthquake early warning have been widely studied. In the field of seismic engineering, ground motions can be mainly classified into near-fault and far-field ground motions. While the ground motion characteristics have a great influence on control performance; however, the existing earthquake early warning system can only predict the peak ground acceleration, and the optimal control efficiency cannot be promptly achieved. Therefore, a prediction module for ground motion characteristics is proposed this study. A database of near-fault ground motions and far-field ground motions is first collected, and the six p-wave features and the high-frequency energy accumulations of the ground dynamic spectrum are used to establish the ground motion characteristic prediction module by utilizing support vector machine. In order to develop the intelligent structural control system, the Leverage-type Stiffness Controllable Isolation System (LSCIS) is used as the structural control mechanism. The effective isolation stiffness of the LSCIS can be swiftly changed to control the dynamic response of the structure. The control parameters corresponding to different types of ground motion are optimized by genetic algorithm, and fuzzy control is adopted for the intelligent isolation system. Finally, performance between the original LSCIS and the proposed intelligent control system is verified.
|
|---|
