This study proposes a Bayesian inference-based decision framework to quantify the physical uncertainty based on fatigue life tests on maraging steel according to post-processing treatments and build orientations. Uncertainty quantification of fatigue life models is performed to determine the most suitable models for the metal additive manufacturing process by employing Bayesian inference. To select one of the fatigue life models, we introduce a weighted-equivalent metric (WEM) to compare the evaluation results from different statistical metrics. By evaluating the WEM value, the logistic model and Zhurkov fatigue life model are identified as the suitable fatigue life models for maraging steel.

本研究提出了一个基于贝叶斯推理的决策框架，以根据后处理处理和构建方向对马氏体时效钢进行疲劳寿命测试，量化物理不确定性。通过采用贝叶斯推理，对疲劳寿命模型进行不确定性量化，以确定最适合金属增材制造过程的模型。为了选择疲劳寿命模型之一，我们引入了加权等效度量 (WEM) 来比较来自不同统计度量的评估结果。通过评估 WEM 值，Logistic 模型和 Zhurkov 疲劳寿命模型被确定为适合马氏体时效钢的疲劳寿命模型。