Published Paper Details:

This Forecasting stream flow in advance is very essential in hydrology such as any water operation and planning. Artificial Neural Network (ANN) recently applied to hydrological related areas. ANN has proven best alternative to convention methods. This study presents the application of artificial neural network to streamflow prediction of stations namely Budhwad in Pune District in Maharashtra, India. In India monsoon occurs only in June to October therefore only June to October data is used for the study. Due to variation in flow in each month therefore separate models are prepared for each month. ANN models developed with four training algorithms namely Levenberg-Marquardt (LM), Conjugate Gradient Function (CGF), Quasi-Newton's back propagation (BFG) and Gradient Descent (GD). The Results from different algorithms are compared with each other. Extreme flow prediction is universal problem to ANN. To overcome this problem one technique is implemented and found to be satisfactory in extreme flow prediction.