Simultaneous estimation of heterogeneous scalar sources and searching for contact-avoidance routes: algorithm construction and its validation

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K03871
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 19010:Fluid engineering-related
• Research Institution: Meiji University
• Principal Investigator: Kametani Yukinori 明治大学, 理工学部, 専任講師 (60759926)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: スカラー源推定 / 乱流 / 随伴最適化 / 直接数値シミュレーション

Outline of Final Research Achievements

In this study, we constructed a framework for the simultaneous estimation of multiple scalar sources and multi-objective path optimization using predicted scalar distributions in a turbulent flow field within an enclosed space. The activities conducted during the period included: (1) constructing a multi-scalar source estimation algorithm using adjoint analysis and verifying it through numerical simulations, (2) employing multi-objective path optimization using reinforcement learning to quickly reach one scalar source while avoiding another specific scalar using the predicted distributions, and (3) experimental validation of the above algorithms. In the source estimation using adjoint analysis, DNS was combined with RANS-based adjoint analysis. For path optimization, we achieved multi-objective optimization by selectively using positive and negative rewards. The experimental validation, however, was limited to constructing an experimental system to obtain local information.

Free Research Field

流体工学

Academic Significance and Societal Importance of the Research Achievements

本研究では火災による被害縮小のため，消防士が局所計測情報から火点・ガス源ならびにその空間分布を予測し，ガスへの接触を最小限にして迅速に火点に到達する最適経路を提示するフレームワーク構築を実施した．随伴最適化による複数種スカラー源推定アルゴリズムは，コロナウイルスを排除しながら効果的に熱を輸送する空調設計や，血流で輸送された病原菌の幹部特定など展開が見込まれる．また，閉鎖空間に限らず大気や海中など開放空間への応用が期待できる．さらに，乱流輸送に関連する諸問題に対し有効な随伴最適化戦略が提示された．経路最適化は今後様々な場面で高度化・実用化が期待されるドローンを用いた極地活動への貢献が見込まれる．