Hasofer-Lind and Rackwtiz-Fiessler (HLRF) method is an efficient iterative algorithm for locating the most probable failure point and calculating the first order reliability index in structural reliability analysis. However, this method may encounter numerical instability problems for high nonlinear limit state function (LSF). In this paper, an improved HLRF-based first order reliability method is developed based on a modified Armijo line search rule and an interpolation-based step size backtracking scheme to improve the robustness and efficiency of the original HLRF method. Compared with other improved HLRF-based methods, the proposed method can not only guarantee the global convergence but also adaptively estimate some sensitive algorithm parameters, such as initial step size, step-size reduction coefficient, using the current known iterative information. Ten selected examples with high nonlinear LSFs are used to compare the robustness and efficiency of the proposed method with the original HLRF method and the improved HLRF (iHLRF) method. Results indicate that the proposed method is not only more computationally efficient but also less sensitive to the remaining user-defined algorithm parameters than the iHLRF method.

Hasofer-Lind和Rackwtiz-Fiessler（HLRF）方法是一种有效的迭代算法，可在结构可靠性分析中找到最可能的故障点并计算一阶可靠性指标。但是，对于高非线性极限状态函数（LSF），此方法可能会遇到数值不稳定性问题。本文基于改进的Armijo线搜索规则和基于插值的步长回溯方案，开发了一种改进的基于HLRF的一阶可靠性方法，以提高原始HLRF方法的鲁棒性和效率。与其他改进的基于HLRF的方法相比，该方法不仅可以保证全局收敛，而且可以自适应地估计一些敏感的算法参数，例如初始步长，步长缩减系数，使用当前已知的迭代信息。选取了十个具有高非线性LSF的示例，以比较该方法与原始HLRF方法和改进的HLRF（iHLRF）方法的鲁棒性和效率。结果表明，与iHLRF方法相比，该方法不仅计算效率更高，而且对其余用户定义的算法参数更不敏感。