Abstract Ground response analysis is greatly affected by uncertainties inherent in soil parameters. However, accurate reliability analysis of the ground response and determination of the reliability index associated with Peak Ground Acceleration (PGA) at the ground surface considering soil layers cross-correlation is very time-consuming. In this paper, by considering two real sites and using frequency domain analysis, the layered-soil sites are regarded as serial components of a cross-correlation system. Then the corresponding reliability index of each is obtained via proposed methods. System reliability indices of the sites PGAs are finally determined by combining the components based on their corresponding correlation through the Sequential Compounding Method (SCM) as a powerful technique for reducing computational costs. The reliability indices obtained by the proposed methods are then compared with the Monte Carlo Simulation (MCS) and Subset Simulation (SS) methods known as efficient MCS. The results show that, in addition to reducing the computational time, the proposed methods represent high accuracy for calculating the reliability indices of ground surface PGAs of the sites.

中文翻译：

---

考虑土层互相关的基于峰值地面加速度的地面响应系统可靠性分析

摘要 地面响应分析受土壤参数固有的不确定性影响很大。然而，在考虑土层互相关的情况下，对地面响应进行准确的可靠性分析并确定与地面峰值地面加速度 (PGA) 相关的可靠性指标是非常耗时的。在本文中，通过考虑两个真实站点并使用频域分析，将分层土壤站点视为互相关系统的串行组件。然后通过提出的方法获得每个相应的可靠性指标。站点 PGA 的系统可靠性指标最终通过基于其相应相关性的组件通过顺序复合方法 (SCM) 进行组合来确定，这是一种降低计算成本的强大技术。然后将通过所提出的方法获得的可靠性指标与称为高效 MCS 的蒙特卡罗模拟 (MCS) 和子集模拟 (SS) 方法进行比较。结果表明，除了减少计算时间外，所提出的方法对于计算站点地表PGA的可靠性指标具有较高的准确性。