论文部分内容阅读
裂隙岩体内的节理往往是变化的,有时甚至变化很大。测得大量节理产状以后,对其分组是一项基础性的工作,但是如何分组目前还没有很好的理论和方法。采用快速聚类分析的方法,将节理产状的样本数据划分为不同的簇,利用极大似然估计的原理,通过数值方法求解费歇尔概率分布模型的参数,并用皮尔逊检验说明了费歇尔逊概率模型的有效性。通过不同簇的概率模型计算及其簇心的分布特性比较说明了簇数分类的适度性。利用所提出的快速聚类分析方法及其推导的计算公式,可以方便地求解费歇尔分布的参数。结合皮尔逊检验,形成了解决这类问题的系统实用的方法。
Joints in fractured rock bodies tend to be variable and sometimes even vary greatly. After measuring a large number of joints, the grouping is a basic work, but there is not a good theory and method to divide the groups. Using the method of fast cluster analysis, the sample data of joints were divided into different clusters. Based on the principle of maximum likelihood estimation, the parameters of the Fischer probability distribution model were solved numerically and the Pearson test was used to illustrate the cost Validity of the Wilson model. By the probability model of different clusters and their distribution characteristics of the cluster centers illustrate the modest number of clusters. Using the proposed fast cluster analysis method and its derived formula, the parameters of the Fischer distribution can be easily solved. Combined with Pearson test, a systematic and practical method to solve such problems has been formed.