摘要
鉴于常用的时频域分析方法在航空电弧故障检测中的局限性,同时为避免单特征量的偶然性,该文提出利用三维熵距的方法实现三种信息熵的特征融合,为减少特征融合带来的噪声与冗余,利用主元分析(PCA)故障检测技术以实现特征矩阵的降维处理。分别分析小波能量熵、功率谱熵、样本熵以及三种信息熵的三维熵距在电弧故障发生前后的特征差异。利用PCA故障检测技术将特征矩阵降维处理,同时给出各试验负载下的PCA的监测统计量T^2和平方预测误差(SPE)及其各自的控制限,通过电弧故障发生前后的T^2和SPE值与各自的控制限的比较,实现航空交流电弧故障的检测。分析结果表明,该方法结合了各特征量的优势,不需要人工设置阈值,能够较为准确地判别电弧故障的发生。最后利用三维熵空间,给出几种典型的电弧故障类型的分类方法。在实现电弧故障检测的同时,给出电弧故障的类型,能够更加有效地对不同电弧故障的发生做出有针对性的应对。
In view of the limitations of common time-frequency domain analysis methods in aviation arc fault detection and to avoid the contingency of single feature,this paper put forward the method of using three dimensional entropy to achieve the feature fusion of three kinds of information entropy.PCA fault detection technology was used to realize dimensionality reduction of the feature matrix for reducing the noise and redundancy of feature fusion.The characteristics of wavelet energy entropy,power spectrum entropy,sample entropy and the three dimensional entropy distance of three kinds of information entropy were analyzed respectively before and after arc faults.The monitoring statistics T^2,SPE and their respective control limits under each test load were acquired by the PCA fault detection technology.The detection of aviation AC arc faults was realized by comparing the values of T^2 and SPE before and after arc faults with their respective control limits.The analysis results show that this method combines the advantages of each characteristic quantity and does not need to set the threshold manually,which can identify the occurrence of arc faults reliably and accurately.Finally,by the three-dimensional entropy space,the classification methods of several typical arc fault types are given.The proposed method provides a reference for realizing the arc fault detection and identifying the type of arc fault.
作者
崔芮华
李泽
佟德栓
Cui Ruihua;Li Ze;Tong Deshuan(State Key Laboratory of Reliability and Intelligence of Electrical Equipment,Hebei University of Technology,Tianjin 300130,China;Key Laboratory of Electromagnetic Field and Electrical Apparatus Reliability of Hebei Province,Hebei University of Technology,Tianjin 300130,China)
出处
《电工技术学报》
EI
CSCD
北大核心
2021年第4期869-880,共12页
Transactions of China Electrotechnical Society
基金
河北省自然科学基金资助项目(E2016202106)。
关键词
航空电弧故障三维熵距
特征融合
PCA检测技术
熵空间
故障分类
Aviation arc fault
three-dimensional entropy distance
feature fusion
PCA detection technology
entropy space
fault classification
