基于TEV、超声波联合检测的开关柜局部放电诊断研究

发布时间：2018-05-10 16:50

本文选题：电力系统 + 高压开关柜　；参考：《华北电力大学(北京)》2017年硕士论文

【摘要】：近年来随着电力系统的快速发展,入网运行的开关柜数量不断增加。但与此同时开关柜故障也逐渐增多,对经济和社会造成巨大损失。因此对开关柜进行不停电检测具有深远意义。TEV法与超声波法的检测技术及特点使其适用于高压开关柜局部放电的检测,论文主要对这两种方法的原理及检测方法进行了深入的研究。在对开关柜进行深入诊断时,所测到的信号是否为局部放电信号是其重要工作之一,通过建立局部放电信号特征提取并建立局部放电典型信号特征库,并建立算法对基于TEV和超声波的原始信号进行特征提取与识别。本文主要对基于TEV、超声波的典型缺陷局部放电特性及模式识别进行了深入研究,详细分析了几种典型缺陷局部放电特性及模式识别方法,并对典型缺陷局部放电的频谱特性、时域分析以及开关柜多源局部放电信号分离技术级应用进行了深入研究。为了深入研究开关柜局部放电特性,分析了开关柜典型局部放电模型。研究发现对于基于TEV法的开关柜局部放电检测,存在多方面问题,不同的放电其放电特性有所不同。同一放电幅值及重复率,对于不同的放电类型,所代表的绝缘状态也可能不同。因此对信号进行放电模式识别是有必要的,结果表明利用放电相位特征可以实现单源放电的模式识别。本文的信号分离与聚类技术不仅能够区分不同放电类型,同时还具有鉴别和分离干扰脉冲的潜力。TEV算法和超声算法两者都是基于连续信号的分析处理,部分算法可以通用,只是相关特征参数不同。基于超声法的模式识别为一般常规信号的模式识别方法,可套用至脉冲电流、超高频、TEV信号的模式识别。最后提出了开关柜局放联合检测方法。通过创新巡检模式,提高巡视质量;分析开关柜运行工况,建立绝缘状态评价体系;多技术联合检测,全面数据综合分析;依据综合评价结果,优化检修策略,准确高效地对开关柜内部局放位置、程度进行深入诊断评估,并结合现场实际检测案例进行分析,证明开关柜联合检测诊断的可行性和有效性。
[Abstract]:With the rapid development of power system in recent years, the number of switchgear running in network is increasing. But at the same time, switchgear failures also gradually increased, causing huge losses to the economy and society. Therefore, it is of far-reaching significance to detect the non-blackout of switchgear. The detection technology and characteristics of TEV method and ultrasonic method make it suitable for the detection of partial discharge of high voltage switchgear. In this paper, the principles and detection methods of these two methods are studied deeply. In the diagnosis of switchgear, whether the measured signal is partial discharge signal is one of its important work. The characteristic of partial discharge signal is extracted and the characteristic library of partial discharge signal is established. An algorithm is established for feature extraction and recognition of the original signal based on TEV and ultrasonic. In this paper, the partial discharge characteristics and pattern recognition of typical defects based on TEV and ultrasonic wave are studied, and several typical defect partial discharge characteristics and pattern recognition methods are analyzed in detail, and the spectrum characteristics of typical defect partial discharge are analyzed in detail. Time domain analysis and multi-source partial discharge signal separation technology for switchgear are studied. In order to study the characteristics of partial discharge of switchgear, the typical partial discharge model of switchgear is analyzed. It is found that there are many problems in partial discharge detection of switchgear based on TEV method, and different discharge characteristics are different. The same discharge amplitude and repetition rate may represent different insulating states for different discharge types. Therefore, it is necessary to recognize the discharge pattern of the signal. The results show that the single source discharge pattern recognition can be realized by using the discharge phase characteristics. The signal separation and clustering techniques in this paper can not only distinguish different discharge types, but also have the potential of discriminating and separating interference pulses. TEV algorithm and ultrasonic algorithm are both based on continuous signal analysis and processing. Only the correlation characteristic parameters are different. The pattern recognition method based on ultrasonic method is a general pattern recognition method for conventional signals, which can be applied to the pattern recognition of pulse current and ultra-high frequency TEV signals. Finally, the combined detection method of switchgear bureau and discharge is put forward. By innovating the inspection mode, improving the inspection quality; analyzing the operating condition of switchgear, establishing the evaluation system of insulation condition; combining multi-technology detection with comprehensive data analysis; optimizing the maintenance strategy according to the comprehensive evaluation results, The location and degree of internal discharge of switchgear are accurately and efficiently evaluated, and the feasibility and effectiveness of joint detection and diagnosis of switchgear are proved.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM591

【参考文献】