Numerical experiment-artificial intelligence approach to develop empirical equations for predicting leakage rates through GM/GCL composite liners

The aim of this study is to develop empirical equations to predict the liquid leakage rate through a composite liner comprising a geomembrane and a geosynthetic clay liner (GM/GCL) underlain by a free draining boundary and having a circular or a longitudinal defect in the geomembrane. For this purpose, an intensive numerical experimental program was conducted where different defect geometries and flow transport characteristics were studied to simulate most of the conditions likely to exist in practice in such type of composite liners. The results are presented in a dimensionless form to generalize the observed behaviour and to give more insight on the factors that control the leakage behaviour. Furthermore, the results are also used to develop empirical equations for predicting the rate of leakage. An artificial intelligent approach referred to as General Method of Data Handling (GMDH) was used for this purpose. The main advantage of the proposed leakage equations is their validity for different flow patterns as the effect of defects geometry and flow characteristics of the composite liner components are already embedded in the development of the equations. However, their validity is limited to the ranges of the dimensionless parameters that were used to develop them.