A Statistical Approach to Confined Seepage Problem with Generation of Random Permeability Field

概要

論文の詳細を見る
In many design practices for rehabilitation or construction of earthen embankments, a deterministic approach is employed for seepage flow analysis, assuming hydrogeological homogeneity of embankment soils. Since in fact soil permeability is heterogeneous, such a conventional approach may often mislead engineers into making an inadequate design with less structural safety and less economical efficiency. This paper intensively investigates reliability of seepage analysis based on a statistical approach that can consider spatial variations of soil permeability. First, random fields with mean, standard deviation and spatial correlation structure are generated, and then finite element seepage analysis in these random fields is performed. The statistical analyses are then implemented through the procedure of Monte Carlo simulation. The statistical solutions to a confined seepage flow problem are compared with the theoretical solution based on deterministic approach. As a result, the mean flow rate Q_m becomes far from the theoretical value Q_t as the global standard deviation σ_k and/or the correlation distance a increase. The spatial correlation structure, not to mention the mean and the standard deviation of permeability, is shown to have a significant effect on seepage flow rate.
日本雨水資源化システム学会の論文