Study On Application Of Artificial Neural Networks To Forecasting Of Ice Jam Water Level

Ice is a common phenomenon in the rivers of cold regions. Rivers in these regions often form the ice cover, the ice jam or the ice dam in winter, and they will bring various ice disasters. The variety of the water level during the period the ice jam is one of the phenomena that forming mostly in the ice jam segment. This article introduced the general situation of the domestic and international ice jam research and the necessity of the water level research firstly. In succession, discussed the formation condition and the formation process of the ice cover in detail, the ice jam formation mechanism and the general regulation of development in the space, the water level variety process that caused by ice jam, factors that influence the ice jam water level: the river power factor, the thermodynamic energy factor, the water current dynamical factor and artificial factor etc.. Then, breifly introduced the arithmetic of the Multiple Linear Regression and artificial neural networks, built up the Multiple Linear Regression model and artificial neural networks model on this foundation, and forecast ice jam water level distinguishly under experiment condition and natural condition. At last, compare results of two methods with measured values. So we can draw a conclusion that the artificial neural networks method can forecast ice jam water level more accurately.