Study Of Fluvial Processes On Yellow River Main And Branch River Based Of Artificial Neural Network

In the past fifty years, the water and sediment process has changed a lot in the Lower Yellow River, resulting in secondary suspended river and main channel shrinkage, which causes great challenge to the river management. The bankfull discharge is significant for potential discharging capacity of the main channel to pass flood flows and to transport sediment. It is one of the core issues to identify the bankfull discharge in the process of governance of the Yellow River. Tongguan is the convergence exports in the Yellow River and Luo River and Wei River, is a natural block port of the Yellow River, is the inflection point of Yellow River in Shanxi- Shaanxi Arear, is erosion base level of North Yellow River and Weihe River. So Tongguan Elevation change has always been concerned. These issues are all key scientific issues in fluvial processes of Middle and Lower Reaches of the Yellow River. In this paper, based on previous studies the author detailedly collected and measured data. The changes and influence factors of Tonguan elevation and bankfull discharge of Lower Yellow River,Lower Weihe River and Inner Mongolia of Yellow River was analysed by using of basic theory of fluvial dynamics and basic principle of fluvial processes. The influence factors of bankfull discharge were divided into there types:â‘ the impact factor of water flow and incoming sediment in flood and non-flood season;â‘¡the accumulative effect of several consecutive years flow discharge and sediment load conditions;â‘¢other factors like peak discharge,sediment content, etc. The influence factors of Tonguan elevation were divided into four types:â‘ the impact factor of water flow and incoming sediment;â‘¡the influence of Three Gorges dam builded;â‘¢the accumulative effect of several consecutive years flow discharge and sediment load conditions;â‘£other factors like peak discharge,sediment content, etc. Artificial neural network is a very effective means to soule complex nonlinear problem. By Artificial neural network, BP models of bankfull discharge and impact factor of bankfull discharge were established in Huayuankou, Gaocun, Aishan, Lijin, Huaxian, Sanhuhekou stations. At the same time BP models of Tonguan elevation and impact factor of Tonguan elevation were established. Through the analysis of model results, the following conclusions was obtained:(1) the impact factor of water flow and incoming sediment in non-flood season can not be ignored. Flood and non-flood water and sediment impact factors are all important to bankfull discharge.(2) The delayed response is a universal phenomenon in the evolution of alluvial river channel. For bankfull discharge calculation, BP method input the data for the past n years can reflect the accumulative effect better. For Tonguan elevation calculation, using Moving average method can get more accurate.(3) Consider other factors such as peak discharge, sediment concentration, sediment particle size is conducive to bankfull discharge increased accuracy. They are the best correlation between Tonguan elevation and the number of days of daily mean flow greater than 2500m~3/s in flood season at Tonguan and Huaxian station. For inflence of operational mode of Three Gorges dam in non-flood season to Tonguan elevation, it has best correlation between Tonguan elevation and the number of days of dam front water level higher than the 315m, 320m and 322m.