论文部分内容阅读
边坡变形预警指标估计对实现边坡安全监控具有重要意义。基于极值理论,建立区间极值(BMM)模型和超阈值(POT)模型,分别利用广义极值分布(GEV)和广义帕累托分布(GPD)对边坡变形监测数据尾部进行拟合分析,结合边坡失事概率,完成对其预警指标的估计。结果表明,两种模型均满足K-S检验,但POT模型可更好地刻画变形监测数据分布的尾部特征,对预警指标的估计偏于安全。研究结果可供类似工程参考。
Slope deformation warning index estimation is of great significance to realize slope safety monitoring. Based on the extreme value theory, interval extreme value (BMM) model and over-threshold (POT) model were established. The tail of slope deformation monitoring data was fitted by generalized extreme value distribution (GEV) and generalized Pareto distribution (GPD) , Combined with the probability of slope failure, completed its early warning indicator estimates. The results show that both models satisfy the K-S test, but the POT model can better characterize the tail characteristics of the deformation monitoring data distribution, and the estimation of the warning index is biased towards safety. The results of the study can be used for similar projects.