基于诱发因素响应分析的滑坡位移预测模型研究

STUDY OF DISPLACEMENT PREDICTION MODEL OF LANDSLIDE BASED ON RESPONSE ANALYSIS OF INDUCING FACTORS

滑坡位移的变化除与其基础地质条件相关之外,更取决于诱发因素的动态作用。为建立滑坡位移动态变化与诱因变化的响应关系,采用时间序列分解预测模型,通过移动平均法将位移分解为趋势项及周期项。趋势项位移由边坡的势能和约束条件所决定,利用多项式位移函数进行拟合预测。周期项位移受库水位涨落和降雨等诱因的周期性动态作用而变化,选取当前月降雨量、累计前两月降雨量、月库水位高程变化量及年内总位移累计增量为影响因子,利用BP神经网络进行多变量位移预测。将各分项位移预测值叠加,从而得到总位移预测值。以三峡库区白水河滑坡为例,利用位移、降雨及库水位变化数据进行计算验证。结果表明,基于滑坡诱发因素和位移变化综合分析预测模型,可以较好地反映诱因动态变化对滑坡位移发展的关键作用,提高预测结果的精度和有效性。

The change of landslide displacement is determined by dynamic functioning of inducing factors besides the basic geological conditions. In order to establish the response relation between dynamic changes of landslide displacement and inducing factors, time series decomposable model is used to decompose the displacement into trend term and periodic term by moving average method. Trend term displacement is determined by the potential energy and constraint condition of the slope and is predicted by displacement polynomial function. Periodic term displacement is affected by the periodic dynamic functioning of inducing factors, such as rainfall, reservoir level fluctuation and so on. The rainfall of current month, cumulative rainfall of anterior two months, reservoir level fluctuation of current month and cumulative increment of total displacement in current year are selected as influencing factors, and the multivariable BP neural network is adopted to predict the displacement. The prediction values of trend displacement and periodic displacement are superposed to obtain total displacement prediction. This model is used to deal with the data of displacement, rainfall and reservoir level fluctuation of Baishuihe landslide in the Three Gorges reservoir area. The results indicate that the prediction model based on the inducing factors and the landslide displacement comprehensively can reflect the key role of dynamic change of inducing factors in displacement development and can improve the precision and effectiveness of prediction results.