SIMULATION OF RAINFALL – RUNOFF PROCESS WITH ARTIFICIAL NEURAL NETWORK (ANN) AND COMPARISON WITH HEC –Ebrahim Afiat Doust, Amirpouya Sarraf and Fardin Boustani
Numerous models have been suggested for illustrate complexity of simulation process of raining to runoff in different studies until now. One of these models is HEC Â– HMS model. Approximately this model is designed in physical mechanisms area which is dominant on hydrological cycles, in fact it is simple form of physical laws and it is shown by parameters indicate basin characteristics. One of the modern methods for simulation process of rainfall to runoff is using and applying artificial neural network. This method which is one of the artificial intelligence methods, it is common because of its nonlinear mathematical structre in area of water engineering sciences. Present study compare above models in simulation process of rainfall to runoff, and it is used from yearly discharge data of Ghareh Aghaj in GharehAghaj basin which is located in south part of Fars province. After completeness of Modeling process and investigations, for choosing network parameters, it is used from error estimating scales which is included MAE1 , RMSE2 , ME3 , GMER4 , GSDER5 and for investigate the correlation between the observing and estimating water measure by model, it is used from R6, SDS 7, SDm 8, SDSD9, LSC10, MSD11 statistics. Finally with consider to in significant difference between two models in this study, the results indicate in compare with HEC Â– HMS model, artificial neural network model shows successful and acceptable results for general simulation of runoff hydrograph.
Enter your contact information below to receive full paper.