基于粗糙集理论的岩体结构面模糊C均值聚类分析  

Fuzzy C-means cluster analysis based on rough set for grouping of discontinuities

作　　者：秦胜伍[1] 陈骏骏 陈剑平[1] 韩旭东[1] 张文[1] 翟健健 刘绪[1] 

机构地区：[1]吉林大学建设工程学院,吉林长春130026

出　　处：《中南大学学报（自然科学版）》2016年第9期3125-3130,共6页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金青年基金资助项目(41202197);国家自然科学基金重点资助项目(41330636);中国博士后科学基金资助项目(20100471265);国土资源部公益性行业科研专项(201211095-6);科技部国家重大仪器科学设备开发专项项目(2011YQ030133)~~

摘　　要：基于在利用模糊C均值聚类算法对岩体结构面产状进行优势分组时,需要人为确定分组数和初始聚类中心,在迭代过程中容易陷入局部最优解的问题,通过改进聚类中心的算法,提出一种基于粗糙集的模糊C均值聚类算法,以优化迭代过程,并通过对比多项聚类有效性检验参数,确定最优聚类分组情况。最后采用模糊C均值聚类算法和改进后的算法对浙江白鹤隧道左洞测得的结构面产状进行优势分组并对比。计算结果表明,本文所提出的方法聚类效果明显优于模糊C均值聚类算法。When using the fuzzy C-means method to analysis the distribution of discontinuities in the discontinuities distribution research in rock mass, the number of group and cluster centers should be firstly determined, which might fall into the locally optimal solution during calculation, in order to solve the problem, a new method was proposed by the optimization algorithm of cluster centers for the dominant partitioning of discontinuities of rock mass based on rough set. This method optimizes the iterative process and can get significant results. Numbers of clustering validity test parameters were taken as laboratory test index to determine the best result of dominant partitioning. Finally, taking the data of discontinuities which were measured in the left tunnel of Baihe in Zhuyong Highway, Zhejiang Province, as an example, the new method and fuzzy C-means cluster was used to analysis and calculate the possible situation. The results show that the proposed method is obviously better than the fuzzy C-means method.

关 键 词：结构面 粗糙集 优势分组 模糊C均值聚类 聚类中心 有效性检验 

分 类 号：P642.3[天文地球—工程地质学]