【摘要】 在实际工程中，时变性与模糊性耦合的现象普遍存在，导致评估系统可靠性时面临模型复杂和计算规模庞大的困难。针对这一问题，从单层解耦结合自适应半代理模型的角度出发，本文提出了模糊失效状态下的时变可靠性分析的高效算法。首先，基于单层解耦将模糊边界通过辅助变量嵌入时变功能函数中，将模糊边界时变可靠性分析等价转换为高维二态假设下的时变可靠性分析；其次，本文借助Kriging代理模型的高斯特性，推导了原始时变功能函数代理模型与等价高维时变功能函数代理模型之间的互推关系，建立了仅代理原始功能函数求解等价高维二态假设下时变可靠性分析的单层半代理模型法；最后，通过数值算例和复合材料机翼模型分析结果验证了所提方法的高效性和准确性。本文所提方法可高效评估复杂工程系统在模糊失效状态下的时变可靠性，为优化和提高系统的安全性与可靠性提供理论指导。更多还原

【Abstract】 In practical engineering, the phenomenon of coupling between time variability and fuzziness is prevalent,leading to difficulties in evaluating system reliability due to model complexity and large computation scales. To address this issue, this paper proposes an efficient algorithm from a single-loop decoupling perspective combined with an adaptive semi-surrogate model. Firstly, the variable that measures the fuzzy property of failure boundary is embedded into the time-dependent limit state function through an auxiliary variable based on single-loop decoupling,which transforms the time-dependent reliability analysis under fuzzy failure states into an equivalent two states timedependent reliability analysis with the relatively high-dimensional variables. Secondly, to avoid establishing the surrogate of the complex equivalent time-dependent limit state function with high-dimensional input variables, the relationship of the surrogate models between the original time-dependent limit state function and the equivalent timedependent limit state function with high-dimensional input variables is derived based on the Gaussian properties of the Kriging model. Then an efficient single-loop semi-surrogate model method for analyzing the time-dependent reliability analysis under fuzzy failure states is constructed. Finally, numerical examples and analyses of composite wing models validate the efficiency and accuracy of the proposed method. This approach enables the efficient evaluation of time-dependent reliability for complex engineering systems under fuzzy failure states, providing a reference for optimizing and enhancing system safety and reliability.更多还原