Groundwater modelling underneath gravity dams considering variable permeabilities and the effect of exit gradients

Abstract

This research focuses on examining the seepage beneath gravity dams built on soil foundations. Unlike typical seepage models, which consider the permeability as a constant within a certain area, this study incorporates the natural variability of properties into the simulation. Specifically, random variables are generated in random fields based on their probability density functions, which include a mean and a coefficient of variation (COV). The implementation is carried out using the Python programming language. The key outcome of seepage analysis for gravity dams is exit gradients which are the main factor causing the internal erosion of foundation in the tailwater side. The outcomes are assessed using Monte Carlo simulations combined with statistical theory. The conclusions should be taken into account when considering internal erosion due to flow gradients; it should be noted that some variable outcomes exceed the deterministic outcome (39,87% larger for exit gradients). This study underscores the effectiveness of stochastic analyses in providing valuable insights into the impact of soil variability, calls for further research in this area, and supports the inclusion of probabilistic techniques in standard design practices.