ニューラルネットワークの履歴推定手法を用いたサブストラクチャ・オンライン実験の精度向上に関する研究

概要

論文の詳細を見る
In general, the substructure On-line test where the hysteresis model is applied for the numerical analysis part cannot simulate dynamic behavior of experiment in real time. The objective of this study is to develop the new experimental technique by estimating the characteristics of numerical part in the substructure On-line test through the neural network. The new learning algorithm to apply a neural network to the substructure On-line test was built, and the normalization method of the new learning data that experiment data to increase in real time can be learned efficiently was proposed and examined about that validity through the numerical analysis. In addition, the numerical experiment, which made the early data of an experiment a parameter was done as a spare analysis of the substructure On-line test by the neural network, and this research discussed the realization and precision of a substructure On-line test by the neural network.
一般社団法人日本建築学会の論文
2004-03-20