并联式加权组合模型在滑坡时间预报中的应用  

Parallel Weighted Combination Model in Time Forecast of Landslide

作　　者：陈巧敏[1] 刘静[1] 侯雪姣 孟婷婷[1] 崔巍[1] 

机构地区：[1]武汉理工大学资源与环境工程学院,武汉430070

出　　处：《武汉理工大学学报》2014年第7期108-113,共6页Journal of Wuhan University of Technology

基　　金：国家自然基金(41171319;40571128;40901214;60904073;41071104);国家重点实验室计划(A0703);863计划(2009AA12Z201);武汉市青年科技晨光计划(200950431203)

摘　　要：针对滑坡预测中单模型不总能进行有效预测且置信度不高的问题,提出了一种新的权重确定方法——并联式权重确定法,用来建立组合模型。证明了预测位移平均相对误差和预测精度之间呈正相关,并对原始Verhulst模型和协同预报模型进行优化。利用新滩滑坡实例和洒勒山滑坡实例进行验证分析。结果表明,基于优化后的Verhulst模型和协同模型的并联式加权组合模型在滑坡中短期时间预报的预测精度远高于原始Verhulst模型和协同模型。A new weighted method-parallel weight determination method,was established to solve the problem that the single model was not always so efficient and confident in time forecast of landslide.Meantime,apositive correlation was proved between the average relative error of monitored displacement and prediction accuracy,which provides theoretical basis for improving the original Verhulst model and synergetic model.These models,including the original and the improved,were applied to the time forecast of some actual landslides that had happened.The forecast results indicated that the parallel weighted combination model based on the improved Verhulst model and synergetic model could make a much more effective prediction than the original single model in time forecast of landslide.

关 键 词：组合模型 并联式权重法 VERHULST模型 协同模型 滑坡预报 

分 类 号：P642[天文地球—工程地质学]