Bayesian Optimised Collection Strategies for Fatigue Testing : Constant Life Testing

This paper presents a statistical framework enabling optimal sampling and robust analysis of fatigue data. We create protocols using Bayesian maximum entropy sampling, which build on the staircase and step methods, removing the requirement of prior knowledge of the fatigue strength distribution for data collection. Results show improved sampling efficiency and parameter estimation over the conventional approaches. Statistical methods for distinguishing between distribution types highlight the role of the protocol in model distinction. Experimental validation of the above work is performed, showing the applicability of the methods in laboratory testing.