Uncertainty Quantification Methods for Helicopter Fatigue Reliability Analysis

The assessment of high-cycle fatigue reliability for helicopter components has been of particular interest to the helicopter community as of late. This reliability analysis typically relies on empirical S/N relationships, Miner's rule for damage summation, and probabilistic models for fatigue strength and flight loads. The probabilistic models for strength and loads are inputs to the reliability calculation, but in practice they are typically estimated on the basis of a limited number of experimental observations or tests. The purpose of this paper is to illustrate a methodology for quantifying the amount of uncertainty associated with reliability predictions when input probability distributions must be estimated from test data. The paper also presents an approach for using probabilistic sensitivity factors to identify the factors that contribute most to the resulting uncertainty. It is expected that these sensitivities can guide decisions to choose the most effective tests (flight tests or component fatigue tests) to reduce uncertainty in the reliability predictions. The AHS Round Robin demonstration problem is used as an example to illustrate these approaches.