LOCAL DAMAGE IDENTIFICATION OF TIMBER FRAME BY NEURAL NETWORK

AIJ
  • JU Dianshu
    Graduate School of Engineering, Kyoto University
  • SUZUKI Yoshiyuki
    Disaster Prevention Research Institute, Kyoto University
  • OU Jinping
    School of Civil Engineering and Engineering Mechanics, Dalian University of Technology

Bibliographic Information

Other Title
  • ニューラルネットワーク手法を用いた木造軸組の局部損傷同定に関する研究

Search this article

Abstract

A neural network approach is proposed for detecting local damages of timber frames based on the fact that the local structural damages will change mechanical properties of the frames. The Elman neural network is selected for the task due to its recurrent connection with feedback from the hidden-layer output to the input. The training data of the neural networks can be obtained from static or dynamic tests of the timber frames. The approach is applied to a full-size traditional timber frame tested on a shaking table. Results show that the neural network approach is effective and accurate to detect different levels of local damages in the timber frame specimen.

Journal

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top