信号峭度不定干扰下的电缆局部放电模式识别研究
Study of Cable Partial Discharge Pattern Recognition under Signal Craggy Indeterminate Interference
高源 1李拥军2
作者信息
- 1. 国网江苏省电力有限公司淮安供电分公司,江苏 淮安 223000
- 2. 海南电网有限责任公司儋州供电局,海南 儋州 571737
- 折叠
摘要
电缆不同类型缺陷引发的放电故障受信号偏斜度、峭度的影响,造成电缆的局部放电信号模式识别效果较差.为此,在信号峭度不定干扰下,提出基于卷积神经网络的电缆局部放电模式识别方法.采用小波阈值去噪方法预处理电缆局部放电信号.运用双树复小波变换提取局部放电信号特征.通过提取偏斜度和峭度刻画信号的分布特征,可以解决信号峭度不定干扰问题.采用卷积神经网络训练提取的特征,并将该特征作为分类器完成电缆局部放电模式识别.试验结果表明,该方法的识别准确率高于95％、识别时间低于5.4 s.该方法应用后具有较好的电缆局部放电模式识别性能.
Abstract
Discharge faults triggered by different types of defects in cables are affected by signal skewness and cragness,resulting in poor partial discharge signal pattern recognition of cables.For this reason,under the interference of signal craggy indeterminacy,the cable partial discharge pattern recognition method based on convolutional neural network is proposed.Wavelet threshold denoising method is used to preprocess the cable partial discharge signal.Dual-tree complex wavelet transform is applied to extract the partial discharge signal features.By extracting skewness and cragness,the distribution of the signal are characterized to solve the problem of indeterminate interference of the signal craggy.Convolutional neural network is used to train the extracted features,and the features are used as classifiers to complete the cable partial discharge pattern recognition.The test results show that the recognition accurancy of this method is higher than 95％and the recognition time is lower than 5.4 s.The method has a good performance of cable partial discharge pattern recognition after application.
关键词
峭度干扰/卷积神经网络/小波阈值去噪/局部放电/双树复小波变换/电缆缺陷Key words
Craggy interference/Convolutional neural network/Wavelet threshold denoising/Partial discharge/Dual-tree complex wavelet transform/Cable defect引用本文复制引用
基金项目
海南省科技攻关计划基金资助项目(183102010257)
出版年
2024
NETL
NSTL
万方数据