Bridge Structural Reliability Analysis Method Based on Kriging Model Optimized by Improved Sparrow Search Algorithm

Bridge engineering as critical transportation infrastructure underpinning national economic lifelines, particularly the service reliability of long-span complex bridge structures, not only determines the system stability of transportation networks, but also directly impacts public safety and property security, and socioeconomic development. However, traditional reliability assessment methods face challenges, such as increased computational demands due to high-dimensional variable coupling, and the difficulty in balancing accuracy and efficiency because of the nonlinear characteristics of the implicit limit state equation. To address these challenges, this study proposes an integrated analytical framework (MISSA-Kriging-MC) combining a Multi-strategy Improved Sparrow Search Algorithm (MISSA), Kriging model and Monte Carlo (MC) method. The proposed method establishes a high-precision response surface of implicit performance functions by the Kriging model, concurrently combining MISSA to optimize the hyperparameter of the Kriging model, which significantly improves the regression capacity and resistance to overfitting of the model; Meanwhile, probabilistic sampling using MC effectively circumvents the complex mathematical solution process and realizes efficient reliability calculation. The feasibility of the method is verified by numerical examples, and on this basis a reliability study is carried out for two large-span bridges. The study demonstrates that MISSA effectively mitigates the inherent limitations of conventional SSA, including premature convergence to local optima and insufficient population diversity, through elite chaotic reverse learning, stochastic following mechanism and Cauchy–Gaussian mutation. Meanwhile, the Kriging model combined with the MC achieves accurate fitting and rapid probabilistic evaluation of high-dimensional nonlinear implicit functions. The method combines high accuracy, high efficiency and strong engineering applicability. It can be easily integrated with the existing finite element analysis software. This integration provides a novel and practical technical pathway for the reliability analysis of complex bridge structures, demonstrating significant potential for engineering applications.