Application of BP Neural Networks to Testing the Reasonableness of Flood Season Staging

Flood season staging scientifically and reasonably can coordinate the relationship of reservoir of flood control and beneficial use, and realize flood resource sustainable utilization. At present, there are many methods of flood season staging, but the results reasonableness testing are remained to be done. In this paper, on the basis of Projection Pursuit (PP) and Set Pair Analysis (SPA) introductions, use the Panjiakou reservoir in Luan River basin as example, divide the flood season with the two methods, and simulate the results with back propagation neural networks (BPNN), use the errors of simulation results as indexes, analyze and contrast the results of the two methods. The result shows that the model is reasonable, it brings up a new idea to test the reasonableness of division of flood season.