The Use of Artificial Neural Networks In The Analysis of Seepage and Slope Stability for The Proposed Qaim Dam on The Khosar River

Abstract

In this study, two Artificial Neural Networks (ANN) were built to predict the amount of seepage and the factor of safety for the upstream and downstream slopes of the Qaim Dam, which was proposed to be constructed on the Khosar River. Two cases have been taken into consideration to analyze the operation of the dam, making use of previous study used Geo-Studio 2012 program studying the stability and the seepage through dam body and its foundation. Thus, two neural networks have been proposed, the first one was for the steady-state case of the reservoir water level and the second was for the rapid drawdown of the reservoir water level. The first ANN gave a coefficient of determination for the seepage process of (0.996),while these coefficients for upstream and downstream slopes were (0.957), and (0.925) respectively. The second ANN deals with calculation of the factor of safety for the upstream slope in a rapid drawdown case, which was (0.976). Sensitivity analyses were also conducted to figure out the most effective variables. It is shown that the most effective factor was the angle of internal friction for the soil.