Monte Carlo模拟和K-均值聚类算法的赤平极射投影概率分析  

The Probabilistic Analysis of the Stereographic Projection Based on the Monte Carlo Simulation and K-means Cluster Algorithm

作　　者：冯登[1,2,3] 王俊杰[1,2,3] 付泓锐 刘云飞[1,2,3] 

机构地区：[1]重庆交通大学国家内河航道整治工程技术研究中心,重庆400074 [2]重庆交通大学水利水运工程教育部重点实验室,重庆400074 [3]重庆交通大学重庆市高校水工建筑物诊断技术与设备工程研究中心,重庆400074

出　　处：《科学技术与工程》2017年第23期76-82,共7页Science Technology and Engineering

基　　金：国家科技支撑计划课题(2015BAK09B01);重庆市基础和前沿研究计划项目(重点)(cstc2015jcyjB X0139);重庆市研究生科研创新项目(20160127)资助

摘　　要：对于结构面多而复杂的岩质边坡,其优势结构面的选取与分析是极其重要的工作。借助于赤平极射投影法,在楔形体滑移分析中首先确定其可能的滑移区域,筛选出可能滑移的结构面交线,缩小计算范围。根据结构面交线的二维经验分布,采用Monte Carlo方法模拟出多条结构面交线,通过结构面交线平均产状的求解来证明该方法的合理性,再采用快速聚类分析的方法,将筛选出的结构面交线划分为不同的组,以赤平极射投影分析得到滑移区域的对称轴中心作为初始凝聚点,通过多次迭代计算得到滑移区域内的优势结构面。将该方法用于重庆綦万高速顺层岩质边坡工程,结果表明,将赤平极射投影与Monte Carlo模拟和K-均值聚类算法相结合,计算得到的优势结构面分类合理,结果可靠,可以准确地确定结构面交线的优势产状。The selection and analysis of the discontinuities are very important works in rock slopes with much complex discontinuities. By using the stereographic projection,the possible sliding zone of wedge can be first determined. The possible sliding intersection lines are selected and the range of calculation is reduced. The intersection lines of discontinuities（ ILD） are simulated by Monte Carlo simulation based on the two-dimensional empirical distribution. The simulation method is proved reasonable by comparing the mean orientation of the sample with the mean orientation of the objective. Joints are categorized into several clusters by using quick clustering analysis. Symmetric axis center of the possible sliding zone is selected as the initial rallying point. The dominant ILD of the possible sliding zone are obtained by iterative calculation. A case study investigated from a highway of the bedding rock slope in Chongqing was used. The study shows that the results are reliable and reasonable. The classification of the ILD and the obtainment of the dominant orientations are more precise.

关 键 词：岩质边坡 赤平极射投影 K-MEANS聚类算法 MONTE CARLO模拟 

分 类 号：TU457[建筑科学—岩土工程]