Investigation of Deep Learning Model for Surface Pressure Estimation by Visualization of Separated Flow around an Airfoil
| Autor: | YAMAMOTO, Yuto, KAWAKAMI, Masashi, RINOIE, Kenichi |
|---|---|
| Jazyk: | japonština |
| Rok vydání: | 2023 |
| Zdroj: | 宇宙航空研究開発機構特別資料: 第54回流体力学講演会/第40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 54th Fluid Dynamics Conference / the 40th Aerospace Numerical Simulation Symposium. :155-165 |
| ISSN: | 2433-2232 |
| Popis: | 第54回流体力学講演会/第40回航空宇宙数値シミュレーション技術シンポジウム (2022年6月29日-7月1日. いわて県民情報交流センター(アイーナ)), 盛岡市, 岩手 The 54th Fluid Dynamics Conference / the 40th Aerospace Numerical Simulation Symposium (June 29 - July 1, 2022. Iwate Citizen Information Exchange Center (Aiina)), Morioka, Iwate, Japan To clarify the separated flow field formed over the airfoil, a Convolutional Neural Network (CNN) model is proposed to obtain surface pressure distributions from visualized images of flow around airfoils. In this paper, we examine the effectiveness of the CNN model by using several datasets with different angles of attack as training data in order to construct a more general-purpose CNN model which can estimate the surface pressure from the visualized images not only at a specific angle of attack but also in a wide range of angles of attack near the stall angle of attack. 形態: カラー図版あり Physical characteristics: Original contains color illustrations 資料番号: AA2230023010 レポート番号: JAXA-SP-22-007 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|