The article present the principle of the ANN briefly in the application of the flood forecast and analyze the shortcoming that the appliance in the flood forecast of the BP method exist and bring forward flood forecast means of the improvement of the BP method. Applying the BP model that possess the delaying function, establishing the delaying neural net work, enhancing the function on ANN and adopting the adding movement means make the web constringency easily. In the researching, we apply MATLAB and VB in the computer formality and exploit the system on the base of the BP in flood forecast. The achievement has been applied in the flood forecasting system of Wenyuhe Reservoir in Shanxi Province. Applying in the practice present that the improved delaying BP means can advance the precision of the forecast and obtain the content precision. And it brings positive effect on the water resource of Wenyuhe reservoir. The application of system has obtained economical and social effect. It is the key problem of ANN modeling that how to confirm networks structure. ANN has been successful research for the application of flood forecast. But there is much study to do in confirm networks structure and building the networks model of much more according with hydrology disciplinarian. The delaying BP algorithm has its characteristic in the design of networks structure. The result applied in the flood forecast of wenyuhe reservoir show the method is effective. The result that the delaying BP algorithm applied in the flood forecast of wenyuhe reservoir show the method is effective.
展开▼