机构地区: 中山大学地理科学与规划学院地理科学系
出 处: 《水文》 1995年第4期6-12,共7页
摘 要: 不直接描述嘉陵江流域三江汇流存在的洪水顶托等复杂的物理过程,而是采用时变参数线性模型,由作者修正的一种自适应滤波水文实时预报方法进行参数在线识别。并分别研究集总输入和多输入两种情况,实现了洪水联机实时预报。结果表明,预报精度较高,且模型简单,滤波参数稳定,便于推广使用。 Not describing the complex mechanism of backwater effect among Jialingjiang River,QujiangRiver and Fujiang River directly,but setting up a time-variant linear model, the paper presents a on-linereal-time hydrological forecasting method.The self-adapting filter algorithm revised by the author is used for theon-line recognition of parameters.The integral input and multiple inputs are studied respectively.The model isapplied to Jialingjiang River Basin for on-line real-time flood forecasting.The results of 65 flood proeesses showthat the model and algorithm are simple and feasible,the parameters and initial values are stable,and the agree-ment between measurements and predictions is good.