基于能量累计曲线-小波变换和动态加权统计的局部放电源定位方法

Location Method of Partial Discharge Sources Based on Energy Cumulative Curve-wavelet Transform and Dynamic Weighting Statistics

为了快速准确地对变压器局部放电(PD)源定位,提出了一种基于能量累计曲线-小波变换和动态加权统计相结合的局部放电定位方法。对获得的超高频(UHF)信号的能量累计曲线进行小波变换来提取超高频信号到达时刻的时间差;由各组时间差数据得到的初始点建立动态加权统计算法的数学模型。排除错误的初始点后,对每个初始点进行动态加权求取PD源位置。结果表明,动态加权统计克服了仅仅根据单组时间差数据定位PD源的分散性,能够更加快速准确地对变压器中的PD源进行定位。使用基于能量累计曲线-小波变换和动态加权统计相结合的方法可以实现对PD源的准确定位,其定位准确度可达10cm。

In order to locate partial discharge（PD） source in transformer rapidly and accurately,we put forward a new partial discharge location method based on the combination of energy cumulative curve-wavelet transform and dynamic weighting statistics.The time delay of ultra-high frequency（UHF）signal,which was extracted from energy cumulative curves of UHF signals,was obtained by using the wavelet transform method.Then a model of dynamic weighting statistics method was established based on the initial points obtained from each time delay data.After excluding initial error points,the PD source of the signals could be located through calculating the dynamic weighting of each initial point.Experimental results show that,the dynamic weighted statistic can overcome the demerit of dispersivity of locating PD source according to a single set of data only,and locate the PD source in transformer more rapidly and accurately.It is concluded that the presented method combining energy cumulative curve-wavelet transform and dynamic weighting statistics realizes accurate location of PD source,with its minimum error of ten centimeters.