三峡库区滑坡土体抗剪强度参数相关性构建的最优Copula函数

Optimal Copula Function Constructed by Correlation of Landslide Soil Shear Strength Parameters in Three Gorges Reservoir Area

Copula函数是一种将联合分布函数与各自边缘分布函数连接在一起的函数,已经成功应用在多个领域.本文为了研究表征抗剪强度参数间相关性的二维Copula函数分布模型最优化识别,在收集的滑坡土体参数基础上,采用K-S检验法确定了抗剪强度参数的最优边缘分布模型,并利用最小平方欧式距离识别出表征抗剪强度参数间相关结构最优的Copula函数.结果表明:Copula函数是构造土体抗剪强度参数c、φ之间联合分布函数的一种有效方法;基于核密度参数估计值的Gaussian Copula函数具有更好地描述参数间相关性的能力,构造方法具有计算简单、适用范围广以及能更好地模拟原始参数间相关性等优点,建议优先采用.该方法应基于监测数据计算后进行Copula函数选择,为工程计算参数选取提供一种新的途径.

Copula function is a connector of joint distribution function and respective marginal distribution function which has been successfully applied to many fields. The optimization identification of two-dimension- al Copula function distribution model representing correlation between shear strength parameters is studied in this paper. First, landslide soil parameters are collected, with the K-S method to determine the optimal mar- ginal distribution model of shear strength parameters. Then, the least square euclidean distances are used to indentify the optimal Copula function representing the correlation structure between shear strength parameters. The results show that Copula function is an effective means to construct the ioint distribution function of soil shear strength parameters Gaussian copula based on kernel density parameter estimated value has better description ability of the parameters correlation; construction method is suggested to use in preference owing to its advantages of simple calculation, wide range of application and better fitting correlation between the original observation data. This method had better to choose Copula functions after calculating the actual measured data, and provide a new way for engineering calculation parameter selection.