摘要
传统的基于局部放电脉冲时频信息构建的局部放电脉冲群分类谱图,多数只能提取表征局部放电脉冲波形特征的低维特征量。当分类算法需要更多的特征量来完成对放电脉冲群的分类工作时,采用上述算法则不能有效地完成对局部放电脉冲群的分类工作。为此提出了采用等效时频熵算法来提取表征局部放电脉冲波形特征的多维特征量,构建放电脉冲群的等效时频熵分类谱图,并与改进的模糊C均值聚类算法相结合实现对不同类型局部放电脉冲群的分类工作。基于气体绝缘组合开关设备(GIS)的实验结果证实了上述方法的有效性和合理性,为研制基于单一人工缺陷模型的局部放电在线监测和识别系统提供了实验和理论依据。
The partial discharge( PD) pulse group classification spectrum constructed based on traditional PD time-frequency information can only provide low-dimensional feature characteristics of PD pluses. When the classification algorithm requires more characteristics of PD pluses to complete the classification work, the abovementioned methods do not work well. This article presents an equivalent time-frequency entropy algorithm to extract the multidimensional characteristics which present the PD pluses waveform feather,and then constructs the PD pluses groups equivalent time-frequency entropy classification spectrum. The spectrum is further combined with the improved fuzzy C means clustering algorithm to complete the classification work of different types of PD pluses groups. The testing results based on gas insulated switches( GIS) prove the validity and rationality of this algorithm,which provides both experimental and theoretical basis for the development of PD online monitoring and identification system based on single artificial defect model.
出处
《电工技术学报》
EI
CSCD
北大核心
2016年第9期181-188,共8页
Transactions of China Electrotechnical Society
关键词
局部放电
模糊C均值聚类算法
特征提取
在线监测
等效时频熵
Partial discharge
fuzzy C means clustering algorithm
feature extraction
online detection
equivalent time-frequency entropy