Design and implementation of materials database for heat treatment process simulation
Zhang Lunfeng, Wang Zhihan, Zhao Junyu, An Kang, Xu Jun, Gu Jianfeng
2023, 48(9):
247-252.
doi:10.13251/j.issn.0254-6051.2023.09.042
Abstract
(
51 )
PDF (3415KB)
(
65
)
Material parameters are the crucial data support in heat treatment process simulation. However, at present, there is a lack of relevant material databases in China, and a few existing databases have problems such as low data accuracy, poor integrity, and inability to share data, they only distinguish materials based on chemical composition, which cannot meet the parameter requirements of heat treatment process simulation. Therefore, a data structure focusing on chemical composition and microstructure was designed, and an online special material database was also independently developed. The database optimizes the data storage structure according to the characteristics of material parameters required for heat treatment process simulation. Adopting B/S architecture design realizes data sharing and improves the convenience of data use. Furthermore, by using data mining technology, the database introduces four machine learning algorithms: multivariable linear regression, Bayesian linear regression, decision tree, and random forest, and establishes an innovative data extraction mechanism. The effective data extraction strategy can be determined through the application analysis of existing data, and then the actual data requires at present can be obtained, which preliminarily solves the problem of data missing currently, and strongly supports the development of heat treatment process simulation.