@inproceedings{oai:jaxa.repo.nii.ac.jp:00003494,
 author = {磯島, 宣之 and 下山, 幸治 and 大林, 茂 and Isoshima, Nobuyuki and Shimoyama, Koji and Obayashi, Shigeru},
 book = {宇宙航空研究開発機構特別資料: 第48回流体力学講演会/第34回航空宇宙数値シミュレーション技術シンポジウム論文集, JAXA Special Publication: Proceedings of the 48th Fluid Dynamics Conference / the 34th Aerospace Numerical Simulation Symposium},
 month = {Dec},
 note = {第48回流体力学講演会/第34回航空宇宙数値シミュレーション技術シンポジウム (2016年7月6日-8日. 金沢歌劇座), 金沢市, 石川, 48th Fluid Dynamics Conference /the 34th Aerospace Numerical Simulation Symposium (July 6-8, 2016. The Kanazawa Theatre), Kanazawa, Ishikawa, Japan, High-resolution turbulent flow simulations using unsteady computational fluid dynamics (CFD) have been widely applied to research and development in aerospace and mechanical engineering industries. In this study, a new data exploration method by sparse structure learning was proposed to detect anomaly elements between two unsteady simulation data sets which have different structures. The new method was tested in unsteady pressure distribution data for two models of an RAE 2822 airfoil with/without a transition trip. The method detected not only obvious change elements such as the transition trip but also small change elements which were easily overlooked by a conventional visualization method, such as delay of turbulent transition., 形態: カラー図版あり, Physical characteristics: Original contains color illustrations, 資料番号: AA1630031019, レポート番号: JAXA-SP-16-007},
 pages = {151--156},
 publisher = {宇宙航空研究開発機構(JAXA), Japan Aerospace Exploration Agency (JAXA)},
 title = {スパース構造学習による二つの非定常流体解析結果データの変化点検出},
 volume = {JAXA-SP-16-007},
 year = {2016}
}