京珠高速K108边坡软弱夹层流变参数的智能反演

Intelligent Back Analysis on Rheological Parameters of Weak Intercalation in K108 Slope of Jingzhu Expressway

岩土体的流变特性是影响边坡变形及稳定性的重要因素,对边坡进行稳定性评价、结合变形监测资料对边坡进行预警预报,都有必要考虑岩土体的流变效应。本文以京珠高速K108边坡为研究对象,采用粘弹塑性数值方法建立该边坡的计算模型,通过大量的数值分析获得神经网络的训练样本,由人工神经网络的非线性映射功能建立岩土体流变位移与待反演参数之间的特征关系,将实测位移代入训练好的神经网络进行反分析得到软弱夹层的流变参数,通过后验差的方法验证了反演结果的合理性,得到的结果可用于后续的边坡稳定分析及预警。

For the slope,the rheological property of rock and soil medium is one of the important reasons for the large deformation and loss of stability.It＇s necessary to take the rheological property into consideration when we evaluate the stability of a slope,forecast the deformation and early warn landslide disaster with monitoring deformation data.This paper back analyzes the mechanical parameters of K108 Slope of Jingzhu Expressway using Artificial Neural Network and Genetic Algorithm.The visco-elastic-plastic computing model of slope is built,and training samples for Artificial Neural Network can be obtained through a lot of calculation.Parameters obtained through back analysis method are verified by posterior variance examination,it＇s proved that the results are reasonable and can be used in subsequent processes.