论文部分内容阅读
由谱系聚类原理,给出了基于各种欧氏距离,即最长距离、最短距离、类平均距离和重心距离的微结构参数间的距离递推公式和微结构参数聚类过程,建立了黏土微结构几何属性的聚类方法。为了评估微结构参数几何各向异性的显著性,给出了各向异性系数的定义。对取自成层地基的某一样品进行了研究,其水平截面、竖直截面和45o截面上的SEM照片二值化参数分析结果表明,根据类平均距离法和重心距离法计算得到的各向异性系数非常接近,因此能很好反映自然固结成层地基的几何各向异性。另外,谱系聚类方法的分析结果与主成分分析方法的分析结果完全一致,这进一步说明,成层地基的几何各向异性是显著的。
Based on the principle of pedigree clustering, the distance recursion formula and the microstructure parameter clustering process based on various Euclidean distances, ie, the longest distance, the shortest distance, the average distance between classes and the centroid distance, are given. Clustering Method of Clay Microstructure Geometric Properties. In order to evaluate the significance of geometrical anisotropy of microstructural parameters, the definition of anisotropy coefficient is given. A sample taken from stratigraphic foundation was studied. The results of binarization parameters analysis on horizontal section, vertical section and SEM photograph at 45o show that according to the average distance method and the centroid distance method, The coefficients of the opposite sex are very close and therefore well reflect the geometrical anisotropy of naturally consolidated layered soils. In addition, the results of pedigree clustering analysis are in good agreement with those of principal component analysis, which further shows that the geometrical anisotropy of layered soils is significant.