Estimation of Seepage Flow Using Optimized Artificial Intelligent Models

Estimation of seepage flow through earthen dam is very important for evaluating stability and safety of the structure. In the present study, we have developed novel optimized artificial intelligent models namely ANN-GA and ANN-BBO which are combinations of Biogeography-Based Optimization (BBO) and Genetic Algorithm (GA) and Artificial Neural Network (ANN) algorithm for the estimation of seepage flow through Karmis earthen dam, Algeria. For the development of models, thirteen years’ period (2006–2018) water level data of the reservoir and dam gallery was used as input data and actual seepage measurement data as output data. Standard statistical measures namely Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and Correlation Coefficient (R) were used to evaluate performance of ANN-GA and ANN-BBO models. Results indicated that performance of both the developed hybrid models is very good but of ANN-BBO model (R: 0.989) is slightly better in comparison to ANN-GA model (R: 0.987) in accurately predicting seepage flow of earthen dam. Thus, proposed optimized artificial intelligence models, especially ANN-BBO can be used for correctly estimating seepage of earthen dams.