利用数据挖掘中的决策树技术,对多年水库实际水文数据和调度数据进行分析和挖掘,从中得到基于多年实际径流资料和水库水位资料的调度决策树模型.模型特点:1)在海量历史水库调度数据中挖掘调度规则,从实践中得到的经验规则往往更具有实用性.2)将入流类别作为判断条件在调度规则中使用,将预报入流类别结合时间、水位生成调度规则更具有合理性.通过实例分析,比较调度决策树模型与常规水库调度图模型的结构特点和调度结果,尝试得到合理的水库兴利调度规则.
Using the technology of decision tree in data mining, a mass of practical hydrological data and reservoir operation information are analysised, and based on this, reservoir operation decision - tree model was established. This model shows notable features and advantages as follow: fist, the reservoir operation rules obtained from masses of historical operation data are more useful and practical than before; second, it takes the classification of forecasting runoff into account and as one of the determining conditions to develop operation rules combined with water - level and real - time, makes the rules more feasible and reasonable. Through comparison of the results attained from regular reservoir operation method and analysis of the operating result of proposed model, the rational operation rules for beneficial purposes are obtained.