灰色关联度法和距离判别分析法在溪洛渡库区潜在滑坡判识中的应用  

Application of grey relation analysis and distance discrimination analysis methods in discriminating potential landslides of Xiluodu Reservoir area

作　　者：李秀珍[1,2] 王成华[2] 邓宏艳[2] 

机构地区：[1]中国科学院山地灾害与表生过程重点实验室,四川成都610041 [2]中国科学院水利部成都山地灾害与环境研究所,四川成都610041

出　　处：《中国地质灾害与防治学报》2010年第4期77-81,共5页The Chinese Journal of Geological Hazard and Control

基　　金：中科院知识创新工程重要方向项目(KZCX2-YW-Q03-5);国家科技支撑计划项目(2006BAC10B04);国家自然科学基金项目(40802072)

摘　　要：以溪洛渡库区深入研究的14个典型滑坡为研究对象,在初步建立潜在滑坡判识指标的基础上,运用灰色关联度理论优选确定了影响库区潜在滑坡孕育的8个关键主控因素:坡面平均坡度、滑体内冲沟切割程度、洼地封闭状况(或排水条件)、滑面平均倾角、滑面形态、下滑段长度与总滑面长度之比、滑体结构及近期活动迹象,并依据距离判别分析的理论和方法,建立了溪洛渡库区潜在滑坡的距离判识模型。该判识模型对库区5个待判滑坡样本和9个滑坡学习样本的判识准确率均达到了100%。Aiming at 14 typical landslides in Xiluodu reservoir area,8 key indexes,including average surface slope angle,cutting degree of the gullies on landslide bodies,closing condition of bottom land（or drainage conditions）,the average inclination of sliding surface,shape of sliding surface,the ratio of length of downslide section and total length of sliding surface,sliding body structure and recent deformation and failure signs,was preferentially gained by using grey relation analysis method,based on preliminarily presented discrimination indexes for potential landslides in that area.Then by using distance discrimination analysis,the discrimination model for potential landslides in Xiluodu reservoir area was built.The results show that prediction accuracy rate of the method is as high as 100%.

关 键 词：灰色关联度法 距离判别分析法 溪洛渡库区 潜在滑坡 判识指标 判识模型 

分 类 号：P642.22[天文地球—工程地质学]