清河水库洪水分类预报方法研究

Study on methods of floodwater classification forecasting in Qinghe Reservoir

采用当前先进的洪水分类预报方法,在流域长系列资料中将非标准洪水提出来,根据其成因不同对其采用人工神经网络分类,利用遗传算法进行参数优选,然后根据洪水特征,选择不同的模型参数和规则进行洪水调度,从而提高水库预报调度精度。

In order to separate non-standard floodwater from long-series data of river basins, advanced floodwater classification forecasting method is applied. Classification of artificial nerve network is adapted according to the cause of flood. Parameter selection is optimized by using genetic calculation method. Then various model parameters and rules are selected for floodwater regulation, so as to increase accuracy of reservoir forecasting and regulation.