滑坡速度倒数法预测模型加速开始点识别及临滑时间预测研究

Research on the identification of acceleration starting point in inverse velocity method and the prediction of sliding time

对矿山或自然土质和岩质边坡而言,大多数滑坡预报都是基于边坡变形三阶段蠕变理论,并根据临滑前加速变形阶段即开始加速点(onset of acceleration,OOA)之后的位移进行滑坡时间预测研究。在分析S-SAR型边坡雷达连续监测的位移后,发现以OOA作为速度倒数法(inverse velocity method,INV)分析的开始点(starting point,SP),所预测的滑坡时间具有一定滞后性。基于变形速度随机变量在斜坡处于匀速变形阶段时服从正态分布特征,提出一种应用正态分布置信区间来动态识别SP位置的方法。通过将S-t坐标系统一量纲后转换成T-t坐标系,建立一种T-lgt的滑坡时间预测模型,此模型应用SP位置后的位移数据可以效提高滑坡预测时间准确性。

For mines or natural soil and rock slopes,most landslide forecast methods are based on the three-stage creep theory of slope deformation,and the prediction of the sliding time is carried out according to the displacement after the onset of the acceleration before sliding. In this paper,after analyzing the displacement continuously monitored by S-SAR slope radar,it is found that the sliding time predicted by using OOA as the starting point of the inverse velocity method has a lag to some extent. Based on the normal distribution characteristics of the deformation velocity random variable in the uniform deformation phase of the slope,a method for dynamic identification of the SP position using the confidence interval of normal distribution is proposed. By converting the S-t coordinate system to the T-t coordinate system,a T-lgt prediction model is established,which can effectively improve the accuracy of landslide prediction time by applying the displacement data after the SP in some case.