基于多维概率分布的砂砾石坝填料级配研究

Gradation of filling materials of sand-gravel dams based on multi-dimensional probability distribution

砂砾石坝填料的级配和粒径决定颗粒间的接触关系,进而影响其宏观力学响应。基于坝体施工过程中检测得到的两千余组试验数据,结合概率统计方法探索颗粒级配与坝体填料基本物理特征间的关系,分别得出了在砂砾石填料区和过渡料填料区与干密度相关性最密切的3个级配特征参数。进而提出基于一维威布尔概率分布模型的坝体填料试验数据分析方法,对以上3个级配特征参数分别构建一维威布尔分布模型,并得到了在一定保证率条件下的粗粒土级配参数估计值。在此基础上,进一步推导出了三维威布尔分布函数表达式,建立了能够利用3个级配特征参数间的相关关系反映干密度的三维威布尔分布模型,采用K-S检验证明了三维威布尔分布模型用于综合考虑多参数影响的坝体填料分析方法的可靠性。研究结果表明,威布尔分布模型能够较好地适用于大坝填料试验数据的描述及参数估计。

The gradation and particle size of filling materials of sand-gravel dams determine the inter-particle contact relationship, and further affect the macro-mechanical response. Based on more than 2000 groups of test data detected during dam construction, the relationship between particle gradation and basic physical characteristics of dam materials is explored by using the probability and statistics method, and three gradation characteristic parameters that are most closely related to dry density in sand-gravel materials and transition materials are obtained respectively. Based on the one-dimensional Weibull probability distribution, the test data analysis method for dam materials is proposed. The one-dimensional Weibull distribution model is respectively established for the above three factors, and the estimated value of gradation parameters of coarse-grained soil is obtained under certain assurance rate. The expression for the three-dimensional Weibull distribution function is further derived, and a three-dimensional Weibull distribution model which can reflect the dry density by using the correlation among the three gradation characteristic parameters is established. The K-S tests are used to verify the fitting degree of the model,which proves the reliability that the three-dimensional Weibull distribution model is used for the analysis of dam materials considering the influences of multiple parameters at the same time.