Prediction of the Seepage Quantity Based on a Random Field Model of Soil Permeability Coefficient

Abstract Natural soils exhibit spatial variability, and predicting seepage quantity considering a random field model of soil permeability coefficient is significant for water conveyance channel. Base on the Yangtze River‐Huaihe River Interbasin Water Transfer Project, employing finite element numerical simulation method to construct a three‐dimensional stochastic field model using normal distribution. Four different stochastic field distribution parameters are researched to predict seepage flow. The main conclusions are as follows: ①Compared to deterministic seepage fields, the average seepage quantity increases about 30% for considering the random distribution of permeability coefficients. ②The selection of parameters in the stochastic field model significantly affects statistical values of seepage flow, with the variance of the fluctuation function having the greatest impact on both the mean and dispersion of seepage flow. Autocorrelation coefficient and fluctuation scale have a significant influence on the dispersion of seepage flow, but little effect on the mean; ③Hydraulic conditions of the slope are crucial factors controlling seepage, and they exhibit clear coupling with the stochastic effects of permeability coefficients. Therefore, in practical engineering, it is necessary to select appropriate hydraulic conditions and stochastic field parameters based on on‐site hydraulic conditions and statistical distribution characteristics of parameters to accurately predict seepage quantity.