An Alternative Probabilistic Approach for Reliability Assessment of Rotorcraft Structures

Abstract

An alternative probabilistic approach is proposed to assess the reliability of rotorcraft structures. According to the approach, safe boundaries of reliability predictions (i. e., conservative ones with additional safety margins) are calculated instead of exact values of the reliability as usually estimated in conventional analysis. Due to additional safety margins, these boundaries are suggested for practical engineering applications. The proposed approach is based on two main ideas, namely (a) prediction of a relatively small population of independent coarse estimations of reliability and (b) application of sampling methods to predict reliability for each individual coarse estimation. Robustness and convenience of the developed approach and its computational implementation are demonstrated for four scenarios considering two problems (simplified analytical and realistic finite element analysis–based ones) under two sets of input data providing the probability of failure (POF) close to 1e-6 and 1e-9, respectively. High accuracy of predictions according to the developed approach is independently verified by comparison with known "exact" references: closed-form solutions for the analytical problem and by direct Monte Carlo simulation for the numerical one. The proposed approach can be recommended as an efficient practical solution for a broad range of reliability assessments of rotorcraft structures requiring both (i) low POF (e. g., below 1e– 6) and (ii) relatively complex time-consuming structural definitions (e. g., compute times in hours for each quasi-deterministic simulation).