Saturated and unsaturated seepage analysis of earth-fill dams using fractal hydraulic conductivity function and its verification

Abstract

One of the most important factors concerning the design of earth-fill dams is the accurate seepage estimation, which is highly correlated to the permeability and hydraulic conductivity of the core material. With the geographic coordinates of 37°31′35.9,“ North and 46°07′55.6″ East, Gale chay dam is located in the northwest of Iran in East Azerbaijan province. The dam is highly important regarding the provided economic benefits for the region. In this research, the steady-state seepage analysis of the Gale Chay dam was conducted via the finite element model (FEM) using a novel saturated fractal hydraulic conductivity model, and the results were compared with those obtained from the conventional and fractal models in the unsaturated condition. For estimating the saturated hydraulic conductivity of the core material, the fractal theory was employed using the image processing technique, the Scanning Electron Microscope (SEM) images, and a new fractal permeability model. The estimated saturated hydraulic conductivities were in good accordance with the measured values. Furthermore, since the fractal theory considers the distribution of the pores, the estimated upper and lower bounds of the saturated hydraulic conductivity resulted in better seepage estimations. Using the estimated maximum, minimum, and average fractal hydraulic conductivities, the results of the saturated seepage analysis indicated relative errors of 15.6%, 8%, and 10.8% compared to the observed seepage. The results were acceptable in terms of safety and economic aspects. Conducting the unsaturated seepage analysis, it was observed that the estimated seepage was highly influenced by the saturated hydraulic conductivity value. Hence, the estimated average fractal hydraulic conductivity by the image processing technique affected the unsaturated fractal and conventional models in a way that the unsaturated seepage analysis was improved by an average of 80%. Moreover, the unsaturated fractal model revealed reliable estimations compared to the conventional models and was considered a feasible alternative for estimating the unsaturated hydraulic conductivity.