论文部分内容阅读
受库水位周期性调度和降雨的影响,三峡库区部分滑坡的位移变形呈台阶状。针对滑坡的这种变形特征,提出一种基于诱发因素响应分析的进化支持向量机位移预测模型:应用移动平均法将滑坡总位移分解为趋势项位移和周期项位移,趋势项位移变化受坡体地质条件控制,应用多项式函数进行预测;周期性位移变化受诱发因素联合控制,选取变形主导因素作为影响因子建立进化支持向量机模型进行预测;将分项位移预测值叠加即为总位移预测值。以库区典型阶跃式滑坡——八字门滑坡为例,应用进化支持向量机模型进行预测研究。结果表明:诱发因素响应分析是滑坡位移预测的关键;基于诱发因素响应的进化支持向量机耦合模型在阶跃式变形期有较好的预测效果,是一种行之有效的阶跃式滑坡位移预测方法。
Due to the periodic regulation of reservoir water level and the rainfall, the displacement and deformation of some landslides in the Three Gorges reservoir area are step-shaped. According to the deformation characteristics of the landslide, this paper proposes an evolutionary prediction model of SVM based on response analysis of evoked factors. The moving average method is used to decompose the total displacement of the landslide into the displacement of the trend term and the term term displacement. The displacement of the trend term is affected by the slope The control of geological conditions, the application of polynomial function prediction; cyclical displacement changes by the joint control of induced factors, select the dominant factor of deformation as an influencing factor to establish the evolutionary support vector machine model to predict; the sub-displacement displacement forecasting value is the total displacement prediction. Taking the step-step landslide-Bazimen landslide in the reservoir area as an example, the evolutionary support vector machine model is applied to forecast the landslide. The results show that the predator response analysis is the key to the landslide displacement prediction. The evolutionary support vector machine coupled model based on the evoked factor response has good predictive effect in the step deformation period and is an effective step-down landslide displacement method of prediction.