Wake-Integral Region Estimation Using Deep Learning for Block-Structured Cartesian Mesh.
-
- NAGAHASHI Shohei
- Kanazawa Institute of Technology
-
- NATSUME Yuta
- Graduate School of Engineering, Kanazawa Institute of Technology
-
- SHIKADA Yusuke
- Graduate School of Engineering, Kanazawa Institute of Technology
-
- SASAKI Daisuke
- Kanazawa Institute of Technology
-
- MATSUSHIMA Kisa
- Toyama University
Bibliographic Information
- Other Title
-
- ブロック構造型直交格子に対する深層学習を用いた後流積分領域推定法
- Part 1: On the Accuracy of Drag Estimation
- その1:抵抗推算精度について
Abstract
<p>Cartesian-Mesh CFD solvers have a problem of inaccurate drag force calculation because the grid does not conform to the object. This is attributed to the fact that near-field drag calculations are performed on the surface of the object, which is represented as a staircase. To avoid this problem, the wake integration method, which calculates the drag from the wake region, has been proposed. However, It is not easy to determine the appropriate integration region. In this research, in order to simplify the selection of the wake region in the drag calculation of the Block-Structured Cartesian Mesh using the wake integration method, deep learning was performed using 70 visualized entropy drag image data. As a result, the prediction worked well for some test data. However, we found that further validation is needed to improve a generalization capability of the network model.</p>
Journal
-
- The Proceedings of Conference of Hokuriku-Shinetsu Branch
-
The Proceedings of Conference of Hokuriku-Shinetsu Branch 2021.58 (0), E011-, 2021
The Japan Society of Mechanical Engineers
- Tweet
Details 詳細情報について
-
- CRID
- 1390289532559211520
-
- NII Article ID
- 130008092491
-
- ISSN
- 24242772
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- Crossref
- CiNii Articles
-
- Abstract License Flag
- Disallowed