Study of Mid and Long-term Runoff Forecast Based on Back-propagation Neural Network

Based on back-propagation (BP) neural network algorithm, by analyzing the data of Dan jiangkou reservoir many years historical runoff series in chronological order and introducing frequency factor, the neural network on mid and long-term runoff forecast has been established. And furthermore, the model has been applied to forecast and analyze the month runoff process of Dan jiangkou reservoir. The case study indicates that the forecasting accuracy of the model has been improved by introducing frequency factor. At the same time, the practical applicability of the model for mid and long-term runoff forecast is verified as well. So, this paper provides a new idea to mid and long-term runoff forecast of reservoirs.