It is pointed out that the Hot Straining Embrittlement (HSE) reduces safety of welded joints. Because, in case that a flaw is detected by NDI techniques and is repaired by renewed welding, the fracture toughness of a flaw not detected but adjacent to the repair part decreases owing to HSE. A reliability analysis is useful to estimate the reduction of the safety related to many deterministic and probabilistic factors. From this view point, the first report described the factors to be considered, the model and the formulations of the reliability analysis, and the probabilities of failure for numerical examples.
In a reliability analysis, it is desirable to investigate sensitivities of parameters to the probability of failure. The sensitivity study, based on the first order and second moment approach, proposed by Shinozuka, is highly effective in order to identify more important parameters through the comparison of sensitivity indexes.
This report presents the results of the sensitivity studies to the probabilities of failure for no repair and repair with or without the Hot Straining Embrittlement.