Multi-layer statistical analysis and surrogate model based fatigue crack growth reliability assessment of lifting lug structure under random load history

The reliability of fatigue crack growth (FCG) in lifting lug structures is a critical component of overall aircraft operational reliability. Assessing the FCG reliability of these structures poses a classic time-dependent challenge, necessitating the consideration of dynamic changes in load distribution, strength distribution, and failure correlations. To address these complexities, a novel FCG reliability assessment method for lifting lug structures has been developed, drawing upon multi-layer statistical analysis theory, artificial neural networks (ANN), and the concept of the minimum order statistic distribution of equivalent strength (Sequ). The introduced variable Sequ enables a direct comparison with applied loads to ascertain structural integrity. Findings suggest that the rate of FCG reliability change is governed by the load and Sequ distributions at any given moment. The predicted Sequ distribution from the new method closely aligns with experimental data, and it offers a more thorough characterization of random load histories. The reliability assessment outcomes of the new method are deemed more rational compared to those of traditional approaches.