Daily Runoff And Suspended Sediment Concentration Forecast For Flood Season Of Tao River Basin

Hongqi station of Tao River is the entrance station of Liujiaxia reservoir. To improve the management both in network and water dispatching and to obtain maximum economic profits, the reservoir operation section of North-west electricity management bureau, the superior of liujiaxia development, decides to set up a system of reservoir automation operation which includes an automatic system of hydrological situation forecasting, a important part of the automation operation system. Taking this practical requirement as study background, this thesis worked on the selection of hydrologic forecasting model and the improvement in forecast precision for the practicing models are complicated in structure, difficulty in multi-parameter calibration, and hard in model updating, for instance, the Xinganjiang model, and the most important thing of all is the fact that there is no sufficient observed data to establish such a model. In addition, this thesis also made an attempt to forecast the hydrograph of suspended sediment for guarantying the safety of water supply from reservoir. The main contents of this thesis and conclusions are shown as follows: 1. Analyze the hydrologic characteristics of Tao River basin, the influence on the hydrology progress imposed by physiographic factors, and variation of the relationship between rainfall and runoff. Identified that the region aboveMinxian station is the key area of runoff-yield2. With the observations and by using SMG (System Model of Genesis) model proposed by Qinyi, the daily runoff hydrograph at Hongqi station of Tao River is forecasted. The results showed that SMG model is simple for applying and not strict with data. The performance of the model had reached the first level according to the special standard, which inferences that the model reflects certain hydrologic behavior. Therefore it is suggested to be applied to practical forecasting.3. Error correction is an important task to real time forecasting. Neural network method was adopted here to build an updating model which predicts the error used in forecast error correction. Proper improvements for BP neural network were also made according to the actual situation. The rectified results show that the determine coefficient of the model has been increased remarkably.4. By analyzing the runoff and sediment characteristics of Tao River basin, it can be concluded that the main sediment-yield region is loess area between Lijia village and Hongqi zone where sediment yield in flood season (June to September) is more than 81 percent of the mean averaged over years.5. Based on the forecasting flow from the real-time forecasting model mentioned above, this thesis tentatively set up the coupling model of rainfall and suspended sediment to serve the water quality control and water supply.