Reliability analysis of UUV pressure hull using improved PSO and adaptive Kriging algorithm

The reliability of pressure hulls critically determines whether unmanned underwater vehicles (UUVs) can successfully accomplish designated missions, rendering accurate reliability assessment indispensable. To enhance the reliability analysis of an unmanned underwater vehicle’s pressure hull, a novel adaptive Kriging model is introduced. This method optimizes the U learning function using an improved PSO algorithm to identify optimal training samples. To refine sample point searches, it combines search space reduction techniques with a distance rule, ensuring precise control over the search region. Additionally, an error-based stopping criterion is implemented to facilitate timely termination. Numerical experiments demonstrate that the proposed method confines the failure probability estimation error within 0.1009% with only 27 limit state function evaluations, significantly improving computational efficiency compared to traditional approaches and reducing the cost of reliability analysis for pressure hulls.