Research on Volume Prediction of Qiantang River Estuary Based on BP Neural Network
Qiantang River estuary is a famous estuary with strong tide and rapid flow,which has a great influence on flood control and drainage.Therefore,higher technical requirements are put forward for the forecast of river volume.Aiming at solving the problem of the low accuracy and convenience of the traditional regression analysis method,a BP neural network model of river bed volume mapping from 4D to 1D was established.Its average error rate and maximum error rate were only 2.65％and 8.76％respectively,which were more reliable and had higher prediction accuracy than the traditional regression analysis method,and it was more convenient to use and update.It can be applied to the long-term and short-term forecast of river volume change,and provide technical support for flood control and drainage of Qiantang River estuary.In addition,using the established neural network model,the relationship between the average flow,the pre-volume and the rate of volume change,as well as the conditions required to maintain the equilibrium volume of the river bed are analyzed.The results show that the pre-volume of 400 million m3 is the critical value for the erosion and sedimentation of the river bed from Zhakou to Yanguan.
river volume predictionBP neural networkQiantang River
万方数据
维普
