平原水库提水调度规则决策树的制定及应用

Application of decision tree in water raising operation regulation in plain reservoir

分析影响提水量的主要因素,以入库泵站运行周期能源单耗最小为目标函数,对历史资料进行优化调度得到水库提水调度数据集,采用决策树C4.5算法从中提取提水调度规则决策树。丁东水库的实例表明,提水调度规则决策树的调度结果相较于实际调度,具有提高供水保证率和泵站效率以及减少能源消耗的优点;相较于遗传算法,具有操作性强的优点。

Main factors affecting the water raising quantity were analyzed in this study. The minimum unit energy consumption of inflow pumps in operation period was selected as the objective function The data set of water raising operation was obtained by optimal operation data. The water raising operation regulation decision tree was extracted from the data set by C4. 5 algorithm. The comparison of operation results by water raising operation regulation decision tree and by actual operation was completed by using Dingdong reservoir as an example. The comparison indicated that the water-supply efficiency and pump efficiency of water raising operation was larger than the actual and energy consumption has the opposite results. Meanwhile, as compared with the genetic algorithm, the water raising operation regulation decision tree has the advantages of strong operability.