Scholarlink

<p>Efficient discovery of advanced additive manufacturing Ni-based superalloys represents an emerging challenge, one that demands balancing minimal manufacturing defects with exceptional material properties. Herein, we propose a framework integrating high-throughput (HTP) approaches and artificial intelligence techniques—specifically HTP thermodynamic calculations, selective laser melting (SLM) experiments, and machine learning (ML)—to develop high-performance SLM-fabricated Ni-based superalloys with reduced printing defects. Guided by fundamental material properties derived from HTP thermodynamic computations and constrained by predictions from ML models, two SLM-fabricated Ni-based superalloys were successfully manufactured. Both alloys exhibit low cracking susceptibility and excellent mechanical properties in the aged state. Notably, their mechanical performance is superior to that of previously reported SLM-fabricated Ni-based superalloys, including IN718, IN939, ADB-850AM, ABD-900AM, and SB-CoNi-10. Leveraging experimental feedback iteratively incorporated into the ML models, we identify that Ni-based superalloys with low Al/Ti contents and high Ta/Mo/Nb contents constitute a promising strategy for mitigating cracking tendency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Rapid Discovery of High-Performance Additive Manufacturing Superalloys Using High-Throughput and Artificial Intelligence Methods

  • 期刊论文

  • 发表日期：

    2025年12月16日

查看全文
  • Pei Liu,
  • Xueting Ren,
  • Shuai Guan,
  • Guangbao Sun,
  • Guohao Liu,
  • Xin Luo,
  • Xiao Sun,
  • Teng An,
  • Zhongnan Bi

摘要

Efficient discovery of advanced additive manufacturing Ni-based superalloys represents an emerging challenge, one that demands balancing minimal manufacturing defects with exceptional material properties. Herein, we propose a framework integrating high-throughput (HTP) approaches and artificial intelligence techniques—specifically HTP thermodynamic calculations, selective laser melting (SLM) experiments, and machine learning (ML)—to develop high-performance SLM-fabricated Ni-based superalloys with reduced printing defects. Guided by fundamental material properties derived from HTP thermodynamic computations and constrained by predictions from ML models, two SLM-fabricated Ni-based superalloys were successfully manufactured. Both alloys exhibit low cracking susceptibility and excellent mechanical properties in the aged state. Notably, their mechanical performance is superior to that of previously reported SLM-fabricated Ni-based superalloys, including IN718, IN939, ADB-850AM, ABD-900AM, and SB-CoNi-10. Leveraging experimental feedback iteratively incorporated into the ML models, we identify that Ni-based superalloys with low Al/Ti contents and high Ta/Mo/Nb contents constitute a promising strategy for mitigating cracking tendency.