一种小波变换和支持向量机相结合的配电网电容器投切扰动源定位方法

A Method to Locate Disturbance Source due to Capacitor Switching in Distribution Network Based on Integration of Wavelet Transform With Support Vector Machine

提出一种小波变换和支持向量机(support vector machine,SVM)相结合的配电网电容器投切扰动源定位方法。该算法使用小波变换对电容器投切产生的暂态扰动信号进行分析,提取扰动能量和扰动持续时间2个特征量作为SVM分类器的输入,以识别扰动,实现投切电容器的自动定位。基于PSCAD/EMTDC仿真软件建立典型10kV配网模型,并选取3个不同位置的电容器进行投切操作,仿真电容器投切产生的暂态扰动。以改变电容器投入时间得到的不同投切事件作为算法样本进行了仿真分析。结果表明,所提算法定位电容器投切扰动源的准确率可达99.17%。

An algorithm to locate disturbance source due to capacitor switching in distribution network, in which the wavelet transform is integrated with support vector machine （SVM）, is proposed. Using wavelet transform, the proposed algorithm analyzes the transient disturbance signal due to capacitor switching and extracts two characteristic quantities, i.e., the disturbance energy and disturbance duration, as the inputs of SVM classifier to recognize disturbances and automatically locate the switched capacitor. Utilizing simulation software PSCAD/EMTDC, a model of typical 10 kV distribution network is built and three capacitor banks at different positions are chosen to be switched, then the transient disturbances due to the switching of capacitor banks are simulated. Taking different switching events obtained by changing the switching-on times of capacitor banks as the samples of the proposed algorithm, the simulation is performed. Simulation results show that the accuracy of locating disturbance source due to capacitor switching can reach to 99.17%.