配网开关用PT在不停电作业工况下铁磁谐振特性与识别研究

Ferromagnetic Resonance Characteristics and Identification of PT for Distribution Network Switchgear Under Uninterruption Working Condition

随着配网自动化技术的推广应用,电压互感器(potential transformer,PT)类设备在搭接操作时若发生铁磁谐振过电压可能导致PT设备过热、开裂或爆炸等故障,威胁作业人员安全。因此,该文针对配网PT不停电作业谐振特性进行分析,采用替代暂态程序-电磁暂态程序(alternative transients program-electro-magnetic transient program,ATP-EMTP)建立10kV不接地系统带电搭接自动化设备的PT铁磁谐振模型,仿真分析了线路长度、PT励磁特性以及PT一次侧中性点接线对配网PT带电作业过程中铁磁谐振的影响,对不同工况下仿真波形采用希尔伯特–黄变换–粒子群优化–遗传算法–支持向量机(Hilbert-Huang transform-particle swarm optimization-genetic algorithm-support vector machine,HHT-PSO-GA-SVM)进行铁磁谐振识别。研究结果表明:在带电搭接PT设备时,当线路的长度满足谐振条件时,其长度增加会使系统依次发生基频谐振和分频谐振过电压;同时,提升PT的励磁特性或PT一次侧绕组中性点接入非线性电阻2种方式可有效抑制铁磁谐振的发生;基于HHT-PSO-GA-SVM的PT谐振识别系统准确率可达92.3%,并有效地识别谐振起始时间及谐振类型。

With the popularization of the distribution automation technology, the installation and replacement of distribution automation switch-gears with PT may involve working on live lines, and the ferromagnetic resonance over-voltage may cause the PT equipment overheating,cracking or explosing, threatening the safety of the workers. In order to understand the resonance characteristics of the PT non-interrupted operation of the distribution network, the ATP-EMTP is used to establish the PT ferromagnetic resonance model for the live connection automatic equipment in the 10kV non-grounded system. The effects of line length, the PT excitation characteristics and the PT primary neutral point wiring on the PT ferromagnetic resonance in the distribution network are studied through simulation calculations. The research results show that as the length of the line increases,the system will have high fundamental frequency resonance and frequency division resonance in succession. When the line increases to a certain length, the PT ferromagnetic resonance phenomenon will disappear. The use of better excitation characteristics or the neutral point of the PT primary winding can effectively suppress the occurrence of ferromagnetic resonance. The PT resonant recognition system based on the HHT-PSO-GA-SVM can reach 92.3% accuracy, effectively identifying the starting time and the types of the resonance.