基于滑坡监测数据的预警判据研究

Research On Early Warning Criterion Of Landslides Using Monitoring Data

位移-时间趋势曲线是研究滑坡失稳预测十分重要的资料依据,它表明了滑坡变形程度及相对的稳定性程度。文章将边坡失稳时的总位移分解成位移的趋势项和位移的周期项,通过时间序列方法对滑坡位移进行预测。文章以德化县上涌镇桂林村马坪滑坡为例,在监测数据的模糊离散性与随机性的基础上以及变形参数所呈现的相关规律,通过GM(1,1)的灰色模型获得位移的趋势项,采用时间序列加法得到滑坡总位移预测值。通过对已获得的滑坡隐患点的监测数据与数值模拟结果的对比分析,建立数学模型,预测滑坡未来的位移趋势,研究滑坡预警判据,最终以位移速率来概化研究区内台风暴雨型滑坡的预警判据。结果显示,该预警判据对于了解边坡位移的发展趋势以及研究边坡的动态稳定性是有效可行的。

Displacement-time curves represent the status and relative stability of landslide of landslide deformation. They are vital to predict landslide deformation monitoring. In this process,the time series analysis method was used to separate the landslide displacement into trend term and seasonal term,the genetic algorithm and support vector regression were adopted to predict the seasonal term displacement.Maping landslide in Dehua for example,based on fuzzy uncertainty and random uncertainty and a certain regularity of deformation parameters,improved grey—time series time-varying prediction model,in which grey theory and traditional time series analysis method are combined. Improved GM（ 1,1） is utilized to extract the trend term of the displacement of slope. Then,the prediction total displacement was obtained by the time series additive mode. The article by using the monitoring data combined with numerical simulation,and mathematical Modeling. the critical displacement rate can be used as the probability threshold interval for warning criterion in landslide area. The model test shows that it is an effective method for prediction of the step type deformation landslides.