Download file (2.09 MB)
Download as zip (1.7 MB)

Collection

Citation

Description:

Creep tests were conducted on weld joints under various welding conditions to investigate the relationship between welding parameters, HAZ shape, and creep performance. The creep life is strongly influenced by stress distribution in the fine-grained heat-affected zone (HAZ), which depends on the HAZ shape. A tandem Bayesian model was developed, combining a Gaussian process (predicting HAZ shape from welding parameters) and Bayesian linear regression (estimating creep life). Welding parameters such as number of layers, layer thickness, and heat input were optimized to find long-life conditions, achieving a creep life 1.8 times longer than the best initial case. Two additional conditions—standard and short-life—were also tested. Creep test results followed the expected trend: long > standard > short. The differences in HAZ shapes corresponded with the differences in creep life. This study demonstrates that creep life can be effectively controlled by optimizing welding conditions.

Rights:

• Creative Commons BY Attribution 4.0 International
  

Keyword: 2 1/4Cr-1Mo steel, Weld joint, Bayesian inference, Creep property, Welding condition, Heat affected zone, Inverse problem

Date published: 2026-01-30

Publisher: Springer Science and Business Media LLC

Journal:

• Welding in the World (ISSN: 00432288) vol. 70 issue. 4 p. 1423-1439

Funding:

• 文部科学省 JPMXP1122684766 (データ創出・活用型マテリアル研究開発プロジェクト（DxMT） 極限環境構造材料研究拠点（RISME）)

Manuscript type: Publisher's version (Version of record)

MDR DOI:

First published URL: https://doi.org/10.1007/s40194-025-02288-9

Related item:

Other identifier(s):

Contact agent:

Updated at: 2026-06-17 10:25:10 +0900

Published on MDR: 2026-06-17 12:39:59 +0900