基于支持度-置信度-提升度的配网自动化系统数据挖掘算法及应用

受到通信或其它设备及配网二次回路故障、缺陷及运行单位人为修改的影响,本已庞大、复杂的配网自动化控制系统数据库的数据质量问题更为严重,数据不准确、不一致、不可靠等问题层出不穷。针对这一问题,文章设计了一种基于支持度-置信度-提升度框架的挖掘算法,从配网自动化控制系统历史数据库中低质量的海量数据中智能挖掘频繁项集,建立符合配网自动化动作逻辑的、具备整体一致性的强关联、强相关规则。文章以具体案例详细介绍了该算法的应用,并通过分析说明该算法不仅为建立客观、合理的配网自动化系统指标评价体系提供可靠、准确的数据挖掘方法以及科学的理论依据,而且可以为管理部门日常运行、管理、维护、消缺工作提供智能的判定工具,节约人力、物力及时间成本,具有很强的工程意义。

a Data mining algorithm to mine the historical big data of the automation operation system of the distribution system based on the Support-Confidence-Lift framework and its application

Affected by the faults, defects of communication equipments and secondary circuits, and artificial records made by operating staff etc., the quality of the mass and complicated datum stored in the database of automation operation system of distribution grids becomes poorer and more worse: incorrect, ununiform, unreliability and so on. To figure out this problem, a data-mining algorithm based on the Support-Confidence-Lift framework is designed to mine frequent itemsets, strong association and correlation rules that conform to the logic of the automation operation system from the massive poor and complicated datum of the automation operation system of distribution networks. In addition, a case on a practical application of the algorithm is studied carefully to illustrate that the designed algorithm can not only offer an exact, reliable data mining method, a theoretical foundation for setting up an objective and reasonable utility indices system, but also provide intelligent judging tools to routine operation, maintenance, defect elimination and management for the management departments of distribution networks, with saving a lot of labor, physical resources and time cost. Therefore, the algorithm is valuable to the engineering practices of distribution grids.