Development And Research Of Flood Forecasting System Of Panshitou Reservoir

According to current obtained hydrologic and phenomena information, flood forecasting forecast a stream cross-section's flow and stage which occur in the future definitive time (named foresee period) . The development reservoir real-time flood forecasting and flood protection scheduling decision support system, implementating the automatic of rainfall regime and water-state information, includeing acquisition, transmission, pretreatment, forecasting, scheduling, named on-line real-time flood forecasting scheduling system, which is the development direction of Water Regime Automatic Surveying System.On the basis of hydrologic characteristic in the Panshitou reservoir, in order to exact and real-time forecast reservoir inflow flood, Sanshuiyuan Xinanjiang distributed model was selected as forecast model, applying Muskingum to proceed stream channel's flow calculation. On the based of VB language and SQL Server 2000 database develop high practicability flood forecast software and implement man-machine conversation function. Software have much function, such as debugging parameter, showing rainfalland flood contrast chart, showing real-time rainfall, inquiry consequence, printing and output etc. The system has been operated in the Panshitou reservoir, which proves that system operation is basical stabilization, flood forecasting precision is fit for the criterion.Because of continuance strengthen complexity and dynamic of hydrology water resources system,continuance enlargemen research scale and influence factor, original research technique and path display definitive difficulty,individual application certainty and uncertainty method face many question that can't be good settle. And so this research apply very popular nervus network that proceed inflow flood's analogue in the Panshitou reservoir in international.Artificial nervus network now is one of the advancing research filed of the world. The research purpose of artificial nervus network is applying software and hardwareto implementation technique, and applying the intelligence information handling ability of artificial nervus network to engineering practice. On the based of the originally method advantage, establishment reasonable artificial nervus network model can settle some complex question that can' t be settled by originally method. In this research, author reference the application, the model regulation and request of some nervus network that be used in flood forecast, which established the nervus network model of suitable for Panshitou reservoir, that is delay nervus network BP model.In this research, author employ two types model to forecast reservoir inflow flood. However in spite of applying which model to proceed flood forecast, system error is inevitable. In this last paper, author analysis andcompare two types model. The consequence show that the forecast runoff result of nervus network BP model is better to the forecast result of Xinanjiang model.