摘要
通过分析地铁供电系统中同一站点的中压开关柜相电流日常规律性特点,利用机器学习方法生成标准曲线。使用3σ准则筛选出粗大误差的异常点值,并根据异常点出现周期性情况进行分类利用。以苏州轨道交通4号线为实例验证,对过去多个周期时段内异常点值出现的频次、设备异常点值数据出现时间范围,从时空斑图上进行比对,研究归纳出异常点值的时空特性。最后,提出基于中压开关柜相电流异常点值的故障预警方法和故障知识库更新策略,从而实现对异常情况及时把控,提高供电设备智能化运维系统的鲁棒性。
By analyzing the daily regularity characteristics of medium voltage switchgear phase current at the same station in metro power supply system,a standard curve is generated by machine learning method.The 3σ criterion is used to filter out the abnormal point values with gross errors,and the abnormal points are classified and utilized according to the periodic occurrence conditions.Taking Suzhou Metro Line 4 as the example for practical verification,the frequency of abnormal point value occurrence and the time range of equipment abnormal point value occurrence in the past multiple cycles are compared from spatiotemporal patterns,and the spatial-temporal characteristics of the abnormal point values are studied and summarized.Finally,a fault early warning method based on the abnormal point value of medium voltage switchgear phase current and an update strategy of fault knowledge base are proposed,so as to realize the timely control of abnormal situation and to improve the robustness of the intelligent operation and maintenance system of power supply equipment.
作者
张锴
曹张保
姜静飞
ZHANG Kai;CAO Zhangbao;JIANG Jingfei(No.1 Operation Branch of Suzhou Rail Transit Group Co.,Ltd.,215100,Suzhou,China)
出处
《城市轨道交通研究》
北大核心
2022年第3期207-212,共6页
Urban Mass Transit
关键词
地铁供电系统
中压开关柜
相电流
异常点
故障预警方案
metro power supply system
medium voltage switchgear
phase current
abnormal point
fault early warning scheme
