Search for high-creep-strength welding conditions considering HAZ shape factors for 2 1/4Cr–1Mo steel

Abstract

The creep rupture life of ferritic heat-resistant steel weld joints is limited by Type IV cracking that occurs in the heat-affected zone (HAZ), whose shape affects creep damage accumulation. In this study, we address the inverse problem of extending the creep rupture life of weld joints by controlling HAZ shape via welding conditions. As reported separately, we have developed a workflow that predicts weld joint creep rupture life from the predicted HAZ shape from welding conditions and have implemented it in the material design system. Using this workflow, we presented a tandem Bayesian model for predicting the creep rupture life from welding conditions via the geometric features of HAZ shapes (HAZ shape factors), which are considered to determine the creep rupture life. The prediction model of a HAZ shape factor from welding conditions was formed by Gaussian process regression. The prediction model of the creep rupture life was formed by Bayesian linear regression. These models were probabilistically connected by Bayesian statistical mathematics. An algorithm to increase the creep rupture life was developed to search for welding conditions. This method was applied to a 2 1/4Cr–1Mo heat-resistant steel weld joint simulated with a plate I-bevel three-layer gas tungsten arc welding. The number of welding conditions combination reaches \({7}^{8}=5764801\). Start from 49 initial HAZ shape factors and 22 creep rupture life data, we performed forward calculations of 20 rupture lives to find welding conditions that can improve the creep rupture life by 12% over the initial data.