Abstract
Abstract In this study, we attempted to model the fatigue life of Ni alloys by employing probabilistic approaches, i.e., survival regressions. The selected models included the Weibull AFT and RSF models. The input variables for the models were the strain amplitude, strain rate, temperature, concentration of dissolved oxygen, and material category (i.e., base or weld). The dataset was divided into two subsets, i.e., training and test sets. The training set was used to train the models, while the test set was used to evaluate the models’ predictions. Several performance metrics were used to measure the model predictabilities, including the c-index, R2, and PCC. Based on performance metrics computed using test sets, all the selected models showed satisfactory performances. The RSF appeared to have better predictability than the AFT model. It was shown that the survival regression models could quantify prediction uncertainty, which is important given the limited fatigue data on Ni alloys.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have