首页|输电线路交叉跨越碰线故障智能预警研究

输电线路交叉跨越碰线故障智能预警研究

扫码查看
为了精准预警输电线路交叉跨越碰线故障,保障输电线路的安全稳定运行,研究一种输电线路交叉跨越碰线故障智能预警方法.设计故障智能预警整体框架,获得目标输电线路交叉跨越碰线故障相关特征数据;采用朴素贝叶斯算法挖掘此类特征数据中的故障因子,并得出其发生指数;利用时间序列一致性故障匹配算法诊断目标输电线路交叉跨越碰线故障,并在碰线故障发生概率较高时发出预警,实现智能预警输电线路交叉跨越碰线故障的目的.测试结果表明:该方法可获得输电线路发生交叉跨越碰线故障时的交、直流线路故障因子电量特征,并在此基础上实现对输电线路交叉跨越碰线故障的智能预警;预警故障 14 次,其准确率为93.3%,预警精度可满足实际应用需求,为输电线路的安全稳定运行提供了保障.
Research on Intelligent Early Warning of Collision Faults at Transmission Line Crossings
In order to provide precise early warnings for transmission line collision faults at the line crossing-over and ensure the safe and stable operation of transmission lines,this paper presents an intelligent early warning method for such faults.The overall framework of the intelligent early warning method is designed,acquiring the collision fault's relevant characteristic data of of the target transmission line.The Naive Bayesian algorithm is utilized to mine the fault factors from these feature data,and to derive their occurrence indices.The time series consistent fault matching algorithm is then used to diagnose the line collision fault of the target transmission line,and give an early warning when the line collision fault probability is high,thus achieving the intelligent early warning of such faults.The test results indicate that the proposed method can obtain the power characteristics of AC and DC line fault factors when a collision fault occurs to the transmission line at the crossing-over,and thus realizing the intelligent early warning:for 14 times of early warning for such faults,the proposed method's accuracy reaches 93.3%,which proves that the early warning accuracy meet the needs of practical application,and provide guarantee for the safe and stable operation of the transmission line.

transmission linescross overwire collision faultintelligent warningnaive Bayesthe time series

王年孝、周华敏、李国强

展开 >

广东电网有限责任公司机巡管理中心,广东广州 510160

输电线路 交叉跨越 碰线故障 智能预警 朴素贝叶斯 时间序列

中国南方电网有限责任公司科技项目

GDKJXM20200432

2024

电网与清洁能源
西北电网有限公司 西安理工大学水电土木建筑研究设计院

电网与清洁能源

CSTPCD北大核心
影响因子:1.122
ISSN:1674-3814
年,卷(期):2024.40(10)