考虑流变的某岩质边坡参数反演分析

Inversion analysis of parameters of rock-slope considering rheological model

针对某水利枢纽工程高边坡长期变形滑动的现象,分析了考虑岩体流变的岩质边坡参数反演问题,依据监测资料,利用神经网络模型建立多输入—多输出非线性拟合模型,得到流变参数—边坡变形量之间的关系;正交实验的方法建立反演分析模型,反演得到H-K流变模型参数。利用反演参数结构计算表明:计算位移与监测变形之间误差在容许范围内。进而利用反演参数进行边坡整体稳定性分析。结果表明:边坡横河向位移变化较大区域发生在高程在1 960 m以上的边坡浅表层区域,区域内最大变形在45~50 mm,在高程1 960 m以下区域位移均较小,主要发生在F42-9断层在坡面出露部位。但整个边坡的流变位移均小于52 mm。

In allusion to rheological model parameters of the rock-slope were gotten through back-analysis method considering the phenomenon of the continued deformation of the high rock-slope.Based on the monitoring data of the slope,the neural network was used to establish the relationship of the multi-input parameters and multi-output displacements.Using orthogonal experiment method,the samples of the calculated values of FEM were gotten.The results of the back-analysis showed that the back-analysis parameters were correct and the error of the calculated deformations and the monitoring data can be ignored.Then the whole deformation of the rock-slope was calculated based on the back-analysis parameters.The results showed that,the maximum of the displacement was occurred upon the 1 960 m part and the value was near 45-50 mm while the deformation blow the 1 960 m was very small and mainly occurred near the F 42-9 fault.All of values of the deformation of the slope were less than 52 mm.