基于关联规则算法的换流站SER事件集挖掘方法

为提高运维人员面对换流站生成的海量事件顺序记录(sequence events recorder, SER)数据的分析能力,提出一种基于关联规则算法的换流站SER事件组挖掘方法。首先利用原始SER事件特征筛选,建立换流站SER事件多维模型;进而利用关联规则算法FP-Growth算法进行数据挖掘与分析,得到换流站典型事件的SER支持组与置信事件;最后基于SER支持组与置信事件分析SER事件集可靠性,方便换流站运维人员及时发现换流站的设备异常动作,减少人工盘查SER造成的事件漏看、错看的可能性。通过挖掘昆柳龙直流(direct current, DC)换流站调试期间SER事件集,表明所提出的方法可以有效地挖掘SER事件集的关联性,为运维人员及时发现SER事件缺失起参考作用。

Association Mining Method for SER Event Sets in Converter Stations Based on Association Rule Algorithm

In order to improve the analysis capability of research and judgment personnel facing the huge amount of sequence events recorder (SER) data generated by the converter stations, an association mining method for SER event sets in converter stations based on association rule algorithm is proposed in this paper. Firstly, the original SER event features are used to filter and build a multidimensional model of the SER events at the converter station. Then, the association rule algorithm FP-Growth algorithm is used for data mining and analysis to get the SER support group and confidence events of typical events in the converter station. Finally, the reliability of SER event set is analyzed based on SER support groups and confidence events, which facilitates the timely detection of abnormal equipment actions in the converter station by the converter station operation and maintenance personnel, and reduces the possibility of missing and misreading events caused by manual inventory of SER. By mining the SER event set during the commissioning of Kun-liu-long DC converter station, the proposed method is shown to be effective in mining the correlation of the SER event set, which serves as a reference for the research and judgment personnel to detect the missing SER events in a timely manner.