Relationship Between High-temperature Plasticity and the Composition of the GH4175 Superalloy Discovered via Machine Learning
Author of the article: LIU Yirui1, LIU Youyun2, ZHAO Jiajun1, HU Xiaobing1, CHEN Yiming1, ZHAO Zhanglong1, LI Junjie1, WA
Author's Workplace:1. State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi'an 710072, China; 2. 93147 Troops of the Chinese People's Liberation Army
Key Words:GH4175; machine learning; high-temperature ductility; component optimization
Abstract:
The GH4175 alloy is a typical new difficult-to-deform superalloy. Optimizing its composition to enhance its
deformation capability in the high-temperature single-phase region is a crucial prerequisite for preventing cracking during
the cogging process of this alloy. By combining thermodynamic calculations, high-temperature tensile experiments, and
machine learning methods, the composition of GH4175 was optimized. The key compositional elements that influence the
γ′ phase volume fraction at 800 ℃, the γ′ phase dissolution temperature, and the alloy melting temperature are identified
via design space screening and adaptive learning strategies. A relationship model between the content of these elements and
high-temperature elongation is established, clarifying the compositional range that ensures excellent high-temperature
ductility while maintaining a reasonable processing window and γ′ phase volume fraction at 800 ℃.