بررسي كارايي روش هاي عصبي - فازي و مدل هاي آماري در شبيه سازي فرآيند بارش - رواناب

Investigation on the efficiency of neuro-fuzzy method and statistical models in simulation of rainfall-runoff process

Rainfall-runoff is one of complex hydrological processes that is affected by a variety of physical and hydrological factors. In this study statistical method ARMAX model, neural network, neuro-fuzzy (ANFIS subtractive clustering and grid partition) and two hybrid models of this methods were used to simulate rainfall-runoff and prediction of streamflow. In each method optimum structure was determined then, streamflow forecasted using the best model. The results showed that hybrid methods have better application than single models and artificial intelligent has better application than linear ARMAX model due to nonlinearity of rainfall-runoff process. In this study all methods showed relatively suitable application but ANFIS method with subtractive clustering is suggested for modeling rainfall-runoff and streamflow prediction.