Reservoir Inflow forecasting Using Artificial Neural Networks (ANNs) in Karkheh Basin

Accurate real-time forecasts of reservoir inflow are specific interest for operation and scheduling in water resources management. There are variety of methods which have been proposed for this purpose including empirical (statistical) and conceptual (physical) approaches. However, this study focused on the utility of Artificial Neural Networks (ANNs) for the forecasting of the daily river flow time series in Kharkheh basin, Iran. The model performance was assessed through MSE, RMSE as well as correlation coefficient. The results indicated that the scenario 3 was found the better than another ones and also showed the lowest error among all scenarios. Therefore the results of the MLP model with different input's scenarios was capable to predict the daily river flow in reservoir inflow of Kharkheh Dam.

Keywords:

Reservoir , MLP , River flow , Karkheh Dam