Development of a Fluid Prediction Model using Rotating Annulus Experiments and Deep Learning and Elucidation of the Typhoon Polygonal Eye Wall Lifecycle

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K03658
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 17020:Atmospheric and hydrospheric sciences-related
• Research Institution: Yokohama National University
• Principal Investigator: Fudeyasu Hironori 横浜国立大学, 教育学部, 教授 (00435843)
• Co-Investigator(Kenkyū-buntansha): 松岡 大祐 国立研究開発法人海洋研究開発機構, 付加価値情報創生部門(情報エンジニアリングプログラム), 副主任研究員 (80543230)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 台風非軸対称構造 / 傾圧不安定 / 回転水槽実験 / 深層学習

Outline of Final Research Achievements

This study focused on the formation of non-axisymmetric structures in typhoons and developed a predictive model combining rotating tank experiments with deep learning. In experiments simulating baroclinic instability, 144 experiments were conducted under varying rotation speeds and water depths to thoroughly investigate the distribution of non-axisymmetric kinetic energy. Velocity analysis using Particle Image Velocimetry and wave number calculation through Fourier analysis were employed. Furthermore, a wave number estimation model using a convolutional neural network, trained with these results, was developed. This model demonstrated high accuracy in 6-class classification and has significant potential to contribute to the prediction of non-axisymmetric structures. These research findings have been widely shared through 5 conference presentations and two journal papers.

Free Research Field

台風

Academic Significance and Societal Importance of the Research Achievements

本研究は台風の非軸対称構造の形成メカニズムを解明し、より正確な台風予測モデルの開発に寄与する可能性がある。特に、回転水槽実験と深層学習を活用した新しいアプローチは、従来の予測手法に比べて精度を大幅に向上させることが期待される。この進展により、台風予測の改善が可能となり、災害対策の効率化や被害の軽減に直接的な影響を及ぼす。社会的な安全と経済的損失の低減に貢献する。また、学術界においては、新たな研究領域を開拓し、気象学の理解を深める契機となる。