桥梁强震资跟工新新程运用总计画.pdf

橋樑強震資料之工程應用─總計畫 羅俊雄 台灣大學土木工程學系 郭鎧紋 黃正耀 中央氣象局地震測報中心 1 2 橋樑強震資料之工程應用─ 子計畫一：應用類神經網路之系統識別 羅俊雄 黃致傑 台灣大學土木工程學系 ABSTRACT A neural-network-based method is proposed to the modeling and identification of discrete-time nonlinear hysteretic system during strong ground motion. The learning or modeling capability of multilayer neural network is explained from the mathematical point of view. The main idea of the proposed neural approach will be explained, and it is shown that multilayer neural network is a general type of NARMAX model and is suitable for the extreme nonlinear input-output mapping problem. Numerical simulation and real structure cases are used here to demonstrate the proposed method. The results illustrate that the neural network approach is a reliable and feasible method. INTRODUCTION The modeling and identification of structural system with nonlinear hysteretic behavior is a widely encountered problem in the field of structural and earthquake engineering. Many efforts have been devoted to establish the models of nonlinear hysteretic system and numerous techniques are developed by many researchers to identify the model parameters. The purpose of these system identification methods is to use numerical simulation or field measurements to improve the dynamic modeling capability for a monitoring of structural systems such as buildings, bridges and dams. Due to the complex nature of civil structures the available response measurement