基于Apriori算法的水库优化调度规则提取方法

Apriori algorithm-based method for extraction of reservoir optimal operation rules

水电站水库长期积累的调度运行数据必然隐藏了一些重要的调度规律,数据挖掘作为从海量数据中自动搜索隐藏于其中的数据规律的技术,具有从水库运行数据中提取隐含调度规则的潜在能力。因此,引入Apriori数据挖掘算法,通过水库优化调度模拟得到长系列调度方案集,进而从调度方案集中找到频繁项集,建立入库径流、发电出力和库水位之间强关联规则,通过此规则制定水库调度方案。将该方法应用于三峡水库年发电优化调度规则提取问题,结果表明,计算得到的年发电量虽然低于优化调度结果,但明显优于常规调度图调度结果,验证了该水库优化调度规则提取方法的合理性和有效性。而且,根据该方法得到的关联规则表结构简单、可操作性强,为解决水库优化调度问题提供了一种新的思路。

Some important scheduling laws are inevitably hidden in the operation scheduling data of hydropower station and reservoir accumulated in a long period. As a technique for automatically searching the data rules hidden in the massive data, data mining has the potential capacity to extract the implicit operation rules from the operation data. Therefore, a long series of scheduling plan sets is obtained through the relevant optimized reservoir scheduling simulation and then Apriori data mining algorithm is used to find out the frequent item sets from the scheduling plan sets for establishing the strong association rules among inflow runoff, power generation output and reservoir water level, thus makes the relevant reservoir scheduling plan through the rules. This method is used to the extract the optimal operation rules for the annual power generation of the Three Gorges Reservoir, from which the result shows that although the calculated annual power generation is lower than that from the result of the optimal scheduling result, it is obviously better than that of the conventional scheduling chart, thus the reasonability and effectiveness of this extraction method of reservoir optimal operation rules are verified. Moreover, the structure of the association rule table obtained according to this method is simple with strong operability, which provides a new idea for solving the problem of reservoir optimal scheduling.