Research Of Runoff Forecasting Based On Artificial Neural Network

Water resource is one of the basic factors that support the sustainable development of life and economics. The variation of the runoff is a primary factor that leads the change of the water system. Affected by the climate, geographic features of the catchment, social development and the necessary needs for the life of human being, the runoff changes complexly.The variation of the runoff has the following features: scale, stochastic, mutation and nonlinearity.Therefore, the prediction of the runoff is one of the research frontiers both in civil and abroad.The thesis based on the runoff of Manas River Kensiwate Hydrological Station, take analysis and forecasting of research.Mature and forecasting methods and new technologies were applied to the runoff forecast in the valley, sum up, research, analysis of a variety of forecasting methods, testing its application and rationality.Manas River flood both the northern slope of Tianshan Mountain snow melt and flood the characteristics of the conditions, but also has its own characteristics.The causes of the flood basin analysis, that the Manas River flood is the main form of snow and floods in the period superimposed on the rainfall over the same penod.They are the dominant form of floods, its characteristics lies in its variability and knowability.To use conventional methods of statistical analysis for the Manas River runoff characteristics,The river runoff changes in the characteristics and features seasonal change during the year to showed;Wavelet analysis of the methods used on the Manas River, the runoff time series there in 27 and 16 around the cycle,They dominated the Manas River, the runoff of long-term changes.And analyze the Manas River, the runoff two main cycle of abundance, litter turn to change the characteristics of volatility.The sunspot activity and the runoff of the basin related analysis, analysis of solar activity on the impact of the runoff;established a number of forecasting model, and compared with results.The paper on the lower reaches of the reservoir not only for optimal operation to provide reliable and reasonable use of the hydrological information,and enhance its economic and social benefits,And sum up and study the model can be widely applied to other watershed runoff forecasts and in used other areas, to provide theoretical and practical reference.