Download file (6.66MB)
Download as zip (5.58MB)

コレクション

引用

説明:

この発表では、人工知能（AI）を用いて提案された非等温時効（NIA）パターンによるNi-Al二元合金（γ–γ'二相）の機械的強度向上に関する解析結果を紹介する。従来の等温時効と比較して、AIにより導かれたNIA条件の一部は優れた0.2％耐力を示した。本研究では計2823通りのNIA条件を評価し、うち173通りが従来法を上回る強度を示した。さらに、AIが発見した条件およびAIに着想を得た専門家設計条件を詳細に解析し、700℃での保持およびその連続回数（二回が最適）、非連続的加熱工程（最大五回）が強度向上に重要であることを明らかにした。

説明:

This study presents the comprehensive analysis of flexible non-isothermal aging (NIA) patterns discovered through artificial intelligence (AI) to improve the mechanical strength (0.2 % proof stress) in γ – γ’ two-phase, binary Ni-Al alloys. In our recent investigation, we found that the AI algorithm could propose aging patterns with superior strength compared to conventional isothermal aging heat treatment. In this current study, we continued our extensive exploration of AI methodologies, uncovering diverse patterns that also surpassed the isothermal aging benchmark. Remarkably, out of 2823 NIA schedules, we found 173 ones outperforming the isothermal aging benchmark. Furthermore, we conducted a detailed analysis of newly AI-discovered patterns and expert-designed patterns inspired by AI. We identified two critical factors for strength improvement: exposure at 700℃ and the number of consecutive 700 ℃ exposures (optimally set at two), alongside non-consecutive steps (up to five). The insights gained from these findings may demonstrate the potential of AI-driven approaches to yield ideas on how to achieve improved strength in Ni-Al alloys.

権利情報:

• Creative Commons BY Attribution 4.0 International
  

キーワード: Artificial intelligence, Monte carlo tree search, Non-isothermal aging, Ni-Al alloys, Pattern analysis

刊行年月日: 2025-02-28

出版者: Elsevier BV

掲載誌:

• Next Materials (ISSN: 29498228) vol. 8 100564

研究助成金:

• JST (The Council for Science, Technology and Innovation (CSTI), Cross-ministerial Strategic Innovation Promotion Program (SIP), "Structural Materials for Innovation" and ”Materials Integration for Revolutionary Design System of Structural Materials.")
• MEXT JPMXP1122684766 (Data Creation and Utilization Type Material Research and Development Project)

原稿種別: 出版者版 (Version of record)

MDR DOI:

公開URL: https://doi.org/10.1016/j.nxmate.2025.100564

関連資料:

その他の識別子:

連絡先:

更新時刻: 2025-11-10 12:30:25 +0900

MDRでの公開時刻: 2025-11-10 12:24:32 +0900