固有モード分解を用いた自動車前面衝突時における車体変形形状の異常検知手法

DOI Web Site

書誌事項

タイトル別名
  • Outlier detection using modal decomposition in vehicle frontal crash analysis

抄録

<p>In recent years, the development process of automobiles has been more complex than ever. It is quite important to properly evaluate the influence on its performance caused by a design change in order to accurately assess the development status and to avoid reworking. However, detecting every important event and so-called outliers during highly complex vehicle crash is a difficult problem. This paper introduces deformation-mode decomposition using principal component analysis and distance measurement in its modal space. With this approach, differences in behavior are represented by the distance between two points in modal space so that the comparisons are objective and quantitative. Besides, subtle changes in the behavior of parts can also be captured compared with conventional approaches using scalar criteria such as distance between two structural points. An example case of this method has been demonstrated with a NCAC full car FE model, and the results with DIFFCRASH and EVENT DETECTION by SIDACT GmbH successfully visualized the change in behavior of the front side frame.</p>

収録刊行物

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ