摘要
使用位移-时间曲线分析及预测滑坡形变趋势时,监测误差和数据录入会影响精度,而恰当的插值方法能提高滑坡位移预测值的准确性。从传统全局分段插值法和滑坡预测经典模型的研究背景出发,充分考虑典型滑坡不同演化阶段对插值效果的影响,将实际位移曲线按实际演化阶段划分为插值算法参数各异的敏感区间,利用多级别阶数的拉格朗日插值法完善滑坡多演化阶段的位移数值填充过程,然后对白家包滑坡的累积位移-时间曲线进行基于实际运动学变化情况的插值,最后通过指数平滑短期预测方法对比验证了该算法对于阶跃型滑坡的预测精度。结果表明,分段插值的方法误差更低,形成的位移预测值更精确,相较于传统方法更具有优势。
When using displacement-time curve to analyze and predict landslide deformation trend,monitoring error and data entry will affect the accuracy,and appropriate interpolation method can improve the accuracy of landslide displacement prediction.Based on the research background of the traditional global interpolation method and the classical landslide prediction model,this paper fully considers the influence of different evolution stages of typical landslides on the interpolation effect,divides the actual displacement curve into sensitive intervals with different interpolation algorithm parameters according to the evolution stages,and uses the multi-level Lagrange interpolation method to improve the displacement numerical filling process of landslide in multiple evolution stages.The cumulative displacement-time curve of Baijiabao Landslide is interpolated based on the actual kinematic changes.Finally,the prediction accuracy of the improved interpolation algorithm is verified by the exponential smoothing short-term prediction method.The results show that the multi-evolution stage interpolation method has lower error,more accurate displacement prediction value,and more advantages than traditional methods.
作者
段功豪
马迅
付杰
DUAN Gonghao;MA Xun;FU Jie(School of Computer Science and Engineering,Wuhan Institute of Technology,Wuhan 430205,China;Hubei Provincial Key Laboratory of Intelligent Robot,Wuhan Institute of Technology,Wuhan 430205,China;Hydrogeological and Environmental Geological Survey Center,China Geological Survey,Baoding 071051,China)
出处
《人民长江》
北大核心
2023年第1期106-110,118,共6页
Yangtze River
基金
湖北省教育厅中青年人才项目(Q20191504)
高等学校学科创新引智计划项目(B17040)
国家重点研发计划重点专项(2019YFC1509605)。
关键词
滑坡预测
多演化阶段
指数平滑
插值算法
landslide prediction
multi-evolution stage
exponential smoothing
interpolation algorithm
