有限要素解析に基づいた構造評価用ニューラルネットワークの構築とコンテナ船の最適設計への利用

概要

論文の詳細を見る
In the structural optimization especially using Genetic Algorithm, the frequency of evaluating individual process is very high, and hence ; simple calculating method are usually used in order to avoid spending huge computational time needed for process. On the other hand, good precision analysis results are also requested. In this report, Neural Network is introduced as alternative method for structure analysis by Finite Element Method. Neural Network has not only advantage but also several weaknesses in its algorithm. We try to back up this failure by adding Lead Factor on the Neural Network system to lead the network to rise high precision analysis. The solution obtained by Transfer Matrix Method is used as lead factor and apply to container ship.
社団法人日本船舶海洋工学会の論文