基于细观统计规律的土石混合体随机模型研究

Research on soil-rock mixture stochastic model based on meso-structural statistical characteristics

对三峡库区一处典型堆积层滑坡的崩坡积土石混合体露头二维照片进行图像处理,获取了土石混合体的细观结构特性;并用最大似然估计法进行概率分布类型拟合,再用Kolmogorov-Smirnov法进行假设检验;根据统计规律建立了土石混合体随机模型。统计结果表明,块体的长轴尺寸与长短轴比服从对数正态分布规律,长轴倾角服从正态分布,分布坐标服从均匀分布,且长轴尺寸的均值和方差均与土石混合体含石量呈线性关系。该随机模型很好地考虑了通常被忽略的小块体,较为细致地反映了土石混合体的细观形态。所建立的土石混合体随机模型可导入有限元软件进行物理力学性质的仿真研究。

Aiming at the colluvial-deluvial soil-rock mixture mass from a typical debris landslide in Three Gorges Reservoir Area, we conduct two-dimensional image process to the picture of outcrop of the mixture, so its meso-structural characteristics are obtained. Then the maximum likelihood estimation is utilized to fit the probability density curve, and then Kolmogorov -Smirnov method is adopted to conduct hypothesis testing. The stochastic model is generated based on the obtained statistical char-acterizations. The results indicate that for the block, the long axis dimension and ratio of long axis to short axis follow the log-normal distribution, while the dip angle of long axis follows normal distribution, and its distributed coordinates follows uniform distribution. Besides, there is linear correlation between the rock proportion in soil-rock mixture and the parameters of mean value and variance of the block long axis dimension. The generated stochastic model includes the small block that was usually ig-nored, so it could reflect meso-structure of soil-rock mixture accurately and by finite element software, the physical and me-chanical characteristics could be analyzed further.