Seismic reliability evaluation of spatially correlated pipeline networks by quasi-Monte Carlo simulation
Seismic reliability evaluation of spatially correlated pipeline networks by quasi-Monte Carlo simulation
复制标题
DOI：
10.1080/15732479.2022.2111445
复制
发表时间：
2022-08
影响因子：
3.7
通讯作者：
B. Hou;Qianyi Xu;Zilan Zhong;Junyan Han;Huiquan Miao;Xiu-li Du
中科院分区：
文献类型：
--
作者：
B. Hou;Qianyi Xu;Zilan Zhong;Junyan Han;Huiquan Miao;Xiu-li Du
摘要自然灾害下的管网可靠性评估有助于提高管网的防灾能力和整体抗灾能力。在地震灾害中，由于管道特性、土壤特性和地震动等关键参数的空间相关性，管网中管道的地震损伤往往具有相关性。提出了一种基于拟蒙特卡罗（QMC）模拟的空间相关参数下管网抗震可靠性评估框架。在明确考虑地面峰值速度、管段壁厚和屈服应力等空间相关随机变量的基础上，研究了空间相关随机变量对管网抗震可靠度的影响。该框架已在两个管网的抗震可靠性评估中得到实施。结果表明，在相同的采样数下，QMC方法比标准MC方法具有更好的模拟精度。空间相关性对管网地震连通可靠性的影响是不同的，管网的拓扑冗余度和从源节点到用户节点的连接路径的平行度不同。对于冗余度和并行度较大的管网，未考虑空间相关性的模型高估了管网连通可靠性。
Abstract The reliability assessment of pipeline networks under nature disasters helps to improve the disaster prevention capability and overall resilience of pipeline networks. For the earthquake hazard, the seismic damage of pipelines in the network is usually correlated due to the spatial correlation of critical parameters, including characteristics of pipeline, soil properties and ground motions. This paper proposes a framework to evaluate the seismic reliability of pipeline network under spatially correlated parameters based on the quasi-Monte Carlo (QMC) simulation. The influence of spatial correlated random variables on the seismic reliability of pipeline network is investigated with explicit consideration of the correlated random variables including the peak ground velocity, the wall thickness and the yield stress of pipe segment. The framework has been implemented in seismic reliability evaluation of two pipeline networks. The result shows that the QMC method has better simulation accuracy than the standard MC method under same sampling numbers. The spatial correlations have different influences on the seismic connectivity reliability of pipeline networks with different topological redundancy and parallelism of connective paths from sources to user nodes. For the pipeline network with greater redundancy and parallelism, the model without considering the spatial correlations overestimates the network connectivity reliability.