Analysis of the staircase method considering the Bayes theorem

The main objective of this paper is to analyze the influence of prior knowledge on the results of the staircase method. In the first part an outline of the application of the Bayes theorem on the classical staircase test procedure is presented. Therefore it is assumed that the fatigue limit is normally distributed. The test data is evaluated by the maximum likelihood method and prior knowledge is integrated by the Bayes method. Confidence limits and parameter calculations are evaluated numerically from the posterior distribution. To analyze the Bayes procedure a Monte Carlo based simulation program was developed. The results of the simulations indicate that the definition of the prior distribution is very important, because it has significant influence on the results.