河道岸线演变预测方法研究

通过分析影响河道岸线变形的因素，选择合适的输入因子，建立预测河道岸线变形的BP神经网络模型，利用闽江竹歧至侯官段的实测资料对网络进行训练学习，然后进行各年份岸线位置的预测，并把预测结果与实测资料进行对比，结果表明，模型预测结果能较好的反映实际岸线的变化情况。

Research on Prediction Method of River Bank-Line Evolution

With analysis of factors influencing the deformation of river bank-line and the selection of some suitable input factors, a BP neural network model is established to predict river bank-line deformation. Based on the measuring materials about the part form Zhuqi to Houguan of Min River, a study about this network is also made to predicate the locations of river bank-line of different years, which in turn are compared with actual measuring data. The research shows that prediction made by the model can properly reflect the changes of the river bank-line.