Estimation of seepage and piezometric pressures on the body of earth dam using soft computing techniques (Case study: Boukan Earth Dam)

Earth dams are always one of the main components of water conservation projects. Nowadays, accurate estimation of piezometric pressure and seepage discharge in earth dams using numerical models and artificial intelligence (AI) approaches is one of the fundamental steps in their design studies. In this research, soft computing models including genetic programming (GP), M5 algorithm and group method of data handling (GMDH) have been used to predict the piezometric pressure in the core and the seepage discharge through the body of Shahid Kazemi Boukan Earth Dam. For this purpose, the information recorded in the last 94 months has been used. The results showed that all of the applied models have permissible level of accuracy in the prediction of seepage discharge and piezometric pressure. The best performance in the piezometric pressure estimation is related to the M5 algorithm with a coefficient of determination (R2) of 0.95 and root mean square error (RMSE) of 0.86. The GMDH by considering the two units (months) delay time and with R2= 0.92 and RMSE=1.541 modeled and predicted the seepage discharge, which was more accurate than other models. In general, increasing the time delay in the input information of models generally increases the performance of proposed models.