Due to the poor application effect of the existing methods in the on-line detection of distribution network,not only the sensitivity is low,but also the specificity is low,so the expected detection effect can not be achieved,this paper presents a nonlinear Support vector machine method for on-line detection of distribution networks.The compression sensing algorithm is used to process the electrical signals of the original distribution network,to reduce the amount of the sample data,to construct the distribution network data series sample set,and to de-noise the data series sample by using the wavelet de-noising technology,by using nonlinear Support vector machine to classify electrical signals and identify abnormal signals,live detection of distribution network can be completed.The experimental results show that the sensitivity and specificity of the design method are above 90％,and the detection accuracy is high,it has a good application prospect in live detection of distribution network.
关键词
非线性支持向量机/配电网/带电检测/压缩感知算法/小波去燥技术
Key words
Nonlinear Support vector machine/Distribution network/Live detection/Compression-aware algorithm/Wavelet desiccant technology
维普
万方数据