Sample Size Reduction Based on Historical Design Information and Bayesian Statistics 2013-01-2440
Numerous test data have been generated in many testing institutions over the years and the historical information from previous similar designs and operating conditions can shed light on the current and future designs since they would share some common features when the changes are not drastic. To effectively utilize the historical information for current and future designs, two steps are necessary: (1) finding an approach to consistently correlate the test data; (2) utilizing Bayesian statistics, which can provide a rigorous mathematical tool for extracting useful information from the historical data.
In this paper, a procedure for test sample size reduction is proposed based on historical fatigue S-N test data and Bayesian statistics. First, the statistical information is extracted from a large amount of fatigue test data collected over the years. Subsequently, the basic theory of the Bayesian statistics and a very effective numerical algorithm for resampling of statistical distributions are introduced. The combination of the historical data and the Bayesian statistics makes the sample size reduction and the assessment accuracy improvement possible. Finally, examples are presented and the benefit of utilizing the new analysis procedure is highlighted and discussed.