摘要
目前对绕组频率响应的分析方法主要是基于幅频响应曲线的分析,因此从频段划分、幅频曲线特征的分析及曲线相似指数3方面综述了幅频响应曲线的研究现状,介绍了频率响应分析法的原理,从扫频范围的选择和影响测量的因素2方面讨论了绕组频率响应的测量,认为应建立原始的绕组频率响应数据库,在反映绕组变形程度的曲线相似指数研究方面有待于提出切实有效的量化标准。
Power transformers are important and expensive units in electric power networks. Short circuit currents or forces during transportation can cause mechanical faults in transformer windings. These faults are winding movement, loss of clamping pressure, buckling, disc deformation etc. Frequency response analysis （FRA） method is more and more frequently being used for identification of transformer windings as a main diagnostic tool. FRA consists of measuring frequency response of transformer windings over a wide range of frequencies and comparing the results with a reference set. This paper briefly introduces the principle of FRA method and presents a review on the current state of FRA method for detecting winding deformation of transformers, including FRA measurements and analyzing techniques and interpretation from the result obtained. FRA measurements is discussed focusing in the selection of swept frequency range and factors that impact the measurements. The FRA results are usually presented on a graph of amplitude or phase against frequency. The phase-frequency graph does not contain as much useful information as the amplitude-frequency graph and it is often not plotted or analyzed. So far the method for analyzing and interpreting the FRA results is based on analyzing and interpreting the amplitude-frequency graph. The current status of analyses of frequency response of transformer windings researches are summarized in regarding the division of frequencies, characteristics of winding frequency response and statistical indicators of the amount of agreement or disagreement between the two sets of measurements. The establishment of FRA results from former times is suggested. There is room for improvement on summation of type of faults in relation to characteristics of the amplitudefrequency graph. An appropriate statistical indication and its criterion should be carefully designed and testified by a large number of experiment data. The future research should emphasize winding deformation location and intelligent diagnosis, such as artificial neutral networks, data mining, or pattern-based fault classification.
出处
《高电压技术》
EI
CAS
CSCD
北大核心
2006年第5期37-41,共5页
High Voltage Engineering
关键词
变压器
绕组变形
频率响应分析法
频谱特性
故障诊断
power transformer
winding deformation
frequency response analysis
frequency spectrum signature
fault diagnosis
