Leakage Rate Prediction through Composite Liner due to Geomembrane Defect using Neural Network

Rakesh Kumar Dutta; Suman Kumari

doi:10.3759/joge.v6i3.3375

Abstract

The paper presents the leakage rate prediction using artificial neural network from the liner made of soil and geomembrane. The defect in the geomembrane considered was having different shapes. Three different shapes such as square, rectangular and circular of the defect have been considered. The input variables considered for the artificial neural network (ANN) were (i) head on the top of the soil (ii) area of the defect (iii) hydraulic conductivity of the soil (iv) thickness of soil and the leakage rate or the discharge was the output. The training of the ANN model was carried out and the weights were obtained. These weights were used to describe the relationship between the input variables and output leakage rate. The study further presented the sensitivity analysis and the variables affecting the leakage rate were identified. Finally, the study concluded that the prediction accuracy using ANN in predicting the leakage rate was quite good.