GIS故障诊断技术研究

发布时间：2018-02-06 06:49

本文关键词： GIS 故障诊断 SF_6气体局部放电最小二乘支持向量机最优权重　出处：《长沙理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：气体绝缘组合电器(简称GIS)具有结构紧凑、可靠性高、安全性好、安装周期短、维护方便、不受外界环境因素影响等优点,在电力系统中得到了广泛的应用及推广。SF6气体作为GIS的绝缘和灭弧介质,利用SF6气体组分能有效诊断及评估其运行状况；局部放电是GIS早期绝缘缺陷的重要表征,能够预测及诊断设备早期故障,避免故障进一步恶化从而影响可靠供电。因此,研究基于SF6气体和局部放电的GIS故障诊断技术具有十分重大的意义。利用SF6气体对GIS进行故障诊断,研究GIS设备运行中温度、风速等环境因素对SF6气体泄漏的影响,提出一种新型的GIS泄漏检测方法,能够在负载电流波动和环境变化时,精确检测SF6开关设备中的气体泄漏率；选取SF6气体分解物组分SOF2/SO2F2、 CF4/CO2、(SOF2+SO2F2)/(CO2+CF4)进行局部放电故障诊断,并使用最小二乘曲线拟合方法得到SOF2、SO2F2气体理想的时间特性曲线及拟合程序,提出采用SOF2、SO2F2两种特征气体可以实现尖端放电、悬浮放电、颗粒放电三种类型的诊断。分析研究GIS尖端放电、悬浮放电、颗粒放电等四种局部放电故障类型,总结出超高频和超声波检测法的放电特征,并进行分析比较,为GIS设备内部放电形式的判断提供依据；采用时域有限差分算法(FDTD)对GIS局部放电时超高频电磁信号进行仿真,并与超高频法(UHF)测得的实际频谱进行比较分析,得出相关结论。基于最小二乘支持向量机的GIS故障诊断方法,建立了LS-SVM模型和分类器,进行GIS尖端放电、悬浮放电、颗粒放电等三种局部放电类型的模式识别；在Matlab中开发基于最小二乘支持向量机的故障诊断程序,并举例论证方法的正确性。基于最优权重的GIS设备综合故障诊断方法,综合超声波检测法、超高频检测法、SF6分解物检测法等三种单一的GIS局部放电检测方法的结果,对尖端放电、悬浮放电、颗粒放电分别进行组合,赋予最优权重,形成新的组合故障诊断方法；采用C语言编写应用程序并进行计算,为基于最优权重的GIS设备综合故障诊断方法提供了良好的操作平台。
[Abstract]:Gas insulated composite appliances (GISs) have the advantages of compact structure, high reliability, good safety, short installation period, convenient maintenance, and not affected by external environmental factors. SF6 gas is widely used in power system as insulation and arc extinguishing medium of GIS. It can be used to diagnose and evaluate the running condition of GIS effectively by using SF6 gas component. Partial discharge (PD) is an important representation of early insulation defects in GIS, which can predict and diagnose early faults of equipment, avoid further deterioration of faults and affect reliable power supply. It is of great significance to study the GIS fault diagnosis technology based on SF6 gas and partial discharge. The SF6 gas is used to diagnose the GIS fault and the temperature in the operation of GIS equipment is studied. This paper presents a new method for detecting SF6 leakage due to the influence of environmental factors such as wind speed and other environmental factors, which can be used to detect the leakage of SF6 when the load current fluctuates and the environment changes. Accurate detection of gas leakage rate in SF6 switchgear; The partial discharge fault diagnosis was carried out by selecting the components SOF _ 2 / so _ 2F _ 2 and CF _ 4 / CO _ 2 / SOF _ 2 SO2F2)/(CO2 _ (CF4) of SF6 gas decomposition. The ideal time characteristic curve and fitting program of SOF _ 2 so _ 2F _ 2 gas were obtained by using least square curve fitting method, and SOF2 was proposed. The fault types of tip discharge, suspended discharge and particle discharge can be diagnosed by two characteristic gases of SO2F2. Four kinds of partial discharge fault types such as tip discharge, suspended discharge and particle discharge of GIS are analyzed and studied. The discharge characteristics of UHF and ultrasonic detection methods are summarized and analyzed and compared to provide the basis for judging the internal discharge form of GIS equipment. A finite difference time-domain (FDTD) algorithm is used to simulate the UHF electromagnetic signals during partial discharge of GIS, and the actual frequency spectrum measured by UHF method is compared with that obtained by UHF method. Based on the least square support vector machine (LS-SVM) method of GIS fault diagnosis, the LS-SVM model and classifier are established to discharge and suspend discharge at the tip of GIS. Pattern recognition of three partial discharge types such as particle discharge; A fault diagnosis program based on least squares support vector machine (LS-SVM) is developed in Matlab, and an example is given to demonstrate the correctness of the method. A comprehensive fault diagnosis method for GIS equipment based on optimal weight is presented. The results of three single GIS partial discharge detection methods, such as ultrasonic detection method, ultra-high frequency detection method and SF6 dissociation detection method, are integrated into tip discharge, suspended discharge and particle discharge respectively. Giving the optimal weight to form a new combined fault diagnosis method; The application program is written in C language and calculated, which provides a good operating platform for the integrated fault diagnosis method of GIS equipment based on the optimal weight.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM595

【参考文献】