Doh, J., Kim, S., Raju, N., Raghavan, N., Rosen, D.W., “Bayesian Inference-Based Decision of Fatigue Life Model for Metal Additive Manufacturing Considering Effects of Build Orientation and Post-Processing,” International Journal of Fatigue, Vol. 155, paper 106535, 2021.  https://doi.org/10.1016/j.ijfatigue.2021.106535

Highlights

  • Investigation of fatigue behavior for metal AM according to the manufacturing process.
  • Uncertainty quantification of fatigue life model for metal AM by employing MCMC.
  • Validation of fatigue life model among candidate models by the u-pooling method.
  • Suggestion of the weighted-equivalent metric (WEM) to decide fatigue life model.
  • Evaluation of goodness of fit of fatigue life models for metal AM using the WEM.

Abstract

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.

The procedure of Bayesian inference-based decision for fatigue life model.