基于数据挖掘的变电站告警数据分析和巡检策略研究

鉴于目前高压变电站告警数据多、运维人员巡检效率低的情况,提出了一种基于关联规则的告警数据分析方法。根据大量告警数据中隐含的先导—后继关系,分析故障发生后产生的告警报文,这有利于找出故障的根本原因;并参考了电网公司的电力设备状态量评定标准,综合告警事件间的关系与各巡检单元权重系数,确定系统频繁告警后巡检的优先顺序,以提高巡检效率。最后,通过某站变压器近一年的告警数据来具体说明这种方法,并且分析了该变压器频繁告警的原因并提出了改进的巡检策略。

Research on Substation Alarm Data Analysis and Patrol Inspection Strategy Based on Data Mining

Considering that the current high-altitude substation"s alarm data is large and the efficiency of operation and maintenance staffs is affected, this paper presents an analysis method of alarm data based on association rules. After identifying the pilot-successive relationship covered in the numerous alarm data, the fault analysis method is constructed, which is helpful for determining the root cause of the fault. Meanwhile, referring to the grid company"s power equipment status assessment standard, alarm events relationships and weights of each inspection unit determine the priority of the inspection after OWS frequently alarms, which contributes to improve inspection efficiency. Finally, this method is illustrated through the alarm data of a station transformer for nearly one year, and the reasons that the transformer frequently alarm are raised. Besides, the inspection strategy is proposed.