Research on PRBP Neural Network Flood Forecasting Model in Chongyang River Basin

The Poak-Ribiére conjugate gradient back propagation algorithm (PRBP) of numerical optimization technology was used, and 21 rainstorm and flood processes from 1997 to 2022 in the upper reaches of Chongyang River basin were studied. The rainfall volume of six rainfall stations in the upper reaches of Chongyang River basin and the previous discharge of Wuyishan Hydrological Station were regarded as input, and its corresponding discharge was regarded as output; the number of hidden layer units was determined by trial calculation, and then PRBP neural network flood forecasting model of Chongyangxi River Basin was established. The remaining eight floods were used to test and validate the model. The results show that compared with that of the conventional BP neural network model, the convergence speed of the model is faster, and the calculation speed is obviously improved; the deterministic coefficient of the model is greater than 0.87, and the relative error of peak flow of six floods is within 10%. The forecasting accuracy meets the requirements, which can provide a basis for the flood control department to forecast the flood.